Healthcare Training Institute - Quality Education since 1979
CE for Psychologist, Social Worker, Counselor, & MFT!!
Target Child Instruction in Initiation Strategies to Promote Interaction
Target Child Initiation Training. Belchic and Harris (1994) suggested that training children with autism to initiate and maintain play with their peers is crucial; such training can overcome the lack of relatedness typical in children with autism and can increase the reinforcement experienced by peer initiators. Interventions to increase the initiations of children with autism typically use some form of prompting and reinforcement to encourage initiations. The peer who is the target of the initiation is told to respond but not initiate. Interventions to increase initiations must take into account spontaneity, success, rate, the appropriateness of behaviors used to initiate, and the demands on the teacher (Zanolli, Daggett, & Adams, 1996). Zanolli et al. used a multiple-baseline-across-activities design in an intervention with two low-functioning 4-year-old children with autism. The intervention consisted of a "priming intervention," in which the two children engaged in their preferred activities and were reinforced for directing teacher-prompted, low-demand social behaviors toward a trained peer. Following priming, the children participated in activity sessions in which teacher prompting was withdrawn and the peer provided reinforcement for initiations. During the intervention phases, the children with autism increased their spontaneous initiations above the average rate of initiations among typical peers without teacher prompt. The initiations were also varied in type, were appropriate, and were successful. This study did not test the generalizability of results to untrained peers.
Belchic and Harris (1994) asked a trained peer to initiate a social interaction with an untrained confederate in the presence of a child with autism. This study involved three preschool-aged children diagnosed with autism and five same-age peers. Prior to the intervention session with the confederate and peer, an adult trainer played with each child with autism and prompted and reinforced initiations. Some adult prompting was also used if the child did not interact with the confederate. Each child with autism reached the criterion of successfully initiating and maintaining social interaction with the confederate at a level equal to or higher than 75% of the time that the confederate spent with the peer. Once criterion was reached with one confederate, the procedure was repeated with a new peer. Less time was required to reach criterion with each subsequent peer. However, the extent of generalization to siblings and on the playground was variable for each child.
Target child initiation interventions have also been used with school-aged children with autism. Using an AB design, Brady, McEvoy, Wehby, and Ellis (1987) employed a loose prompting technique in which an adult prompted an 11-year-old child with autism to play with peers during a training session. The target child's spontaneous initiations increased with both trained and untrained peers. Gunter, Fox, Brady, Shores, and Cavanaugh (1988) used a similar approach, whereby the teachers prompted two children diagnosed with autism to initiate with peers, who were instructed to respond to the target child's initiations. Using a multiple-baseline design, they found that both participants increased their initiations and interactions when teacher prompts and praise were directly applied. However, only one child generalized to untrained peers in a new setting without teacher prompt.
Initiation Training for Target Child and Peers. Gonzalez-Lopez and Kamps (1997) used a strategy for increasing the interactions of children with autism that combined several approaches. First, peers were given information about disabilities and were taught behavior management skills, including giving easy instructions, prompting, reinforcing, and ignoring disruptive behaviors. The peers and the children with autism were then taught how to use greetings, how to initiate play, conversation strategies, imitation, how to follow instructions, how to share, turn taking, how to ask for help, and how to request things. During the last condition, teachers reinforced interactions by using a star chart. All intervention conditions consisted of 20-minute small play groups. A multiple-baseline design was used to evaluate this intervention with four children with autism ranging in age from 5 to 7 years who had at least moderate comprehension and some limited verbal communication. With the exception of one child, the mean frequency of interactions increased from 2.3 to 3.4 with social skills training and from 6.6 to 8.3 with social skills training plus reinforcement. In addition, behavior problems were reduced, which often increases the likelihood that peers will persist in interactions with children with autism. There was evidence, however, that the more disruptive children might require additional behavior management strategies.
Comparative Studies of Various Approaches
Oke and Schreibman (1990) investigated the peer-initiation training strategy as well, comparing it to other peer training strategies and a target child training methods. Using a multiple-treatment design with reversal phases, one high-functioning 5-year-old child with autism, two trained 4- and 5-year-old peers, and one untrained 7-year-old peer participated in three different interventions. Intervention 1 involved peer-initiation training with sharing and play organizing strategies, as previously discussed. Intervention 2 also used this approach; however, peers received additional instruction in discrimination between parallel and interactive play. Intervention 3 focused on training the child with autism; the intervention included training in sharing, play organizing, and turn taking, similar to that given to peers in Intervention 1. Training the peer in initiation strategies (Intervention 1) led to increases in social interaction, but such interaction was highly variable. Intervention 2 produced higher and more stable interaction, but interaction dropped dramatically during a reversal phase in which peers no longer used the initiation strategies. After Intervention 3 was instituted, in which the child with autism was trained to initiate, the increases in interaction were maintained during a subsequent reversal phase. Interventions 2 and 3 both led to higher rates of interaction than Intervention 1, but they did not significantly differ from each other. Notably, the target child's affect changed from neutral to positive only during Intervention 3. There was some evidence for generalization to an untrained peer, though generalization did not occur during free-play time at school.
Important Issues in Socialization Interventions
Though all of the reviewed studies reported substantial improvements in social interactions, the nature of the improvements varied substantially, both across studies and between participants. Frea (1995) noted that an important difficulty in producing stable and substantial improvements with socialization interventions for children with autism is that generally the interventions can address only a small facet of the problem, which leads to negligible changes in the amount or quality of social interaction. The social validity of the increases in interaction produced by peer-mediated interventions is also a concern. Many studies reported increases in responding or initiation by target children, but not both; yet both are required for reciprocity. In addition, though most studies reported increases in interactions, the results are rarely compared to the level at which typical children interact (Pollard, 1998), and there is a strong emphasis on quantitative measures rather than qualitative ones (McEvoy & Odom, 1987). Reliance on quantitative measures ignores the fact that often the interactions of children with autism remain idiosyncratic. The quality and complexity of interactions must also be considered if interventions are to truly affect naturally-occurring peer relationships (Mundschenk & Sasso, 1995).
Though peer-mediated interventions have shown promise and deserve further attention, maintenance and generalization of improvements in social interaction have been less pronounced (McEvoy & Odom, 1987; Kohler et al., 1997; Krantz, 2000). These are critical issues because social skills are necessary at most times, in all contexts, with most people (Kohler et al., 1997). Generalization is important across both settings and peers.
Setting events are important determinants of social behavior (Gaylord-Ross & Haring, 1987). Honig and McCarron (1988) confirmed the importance of setting for generalization. In their study of preschool children in unstructured playtime, circle time, structured play, and gym, the frequency of the interactions of children with autism varied by setting. In addition, McEvoy and Odom (1987) noted that the type of activity also influenced the child's interactional behavior. Furthermore, interactional skills learned with the assistance of one peer may not be used spontaneously with other peers (Belchic & Harris, 1994). Breen, Haring, Pitts-Conway, and Gaylord-Ross (1985) proposed the multiple-peer tactic, with the assumption that generalizability will increase if children are given the opportunity to practice with several peers who have different response characteristics. Bandura (1977) also suggested that diversity in modeling fosters behavior acquisition.
Though the use of multiple peers has been recommended as useful for increasing generalization, the impact on peers remains an important consideration in peer-mediated interventions. Peer response is especially important if a change in peer expectancies is necessary to promote interaction with children with autism. Peck, Donaldson, and Pezzoli (1990) addressed the impact of socialization interventions on peers through interviews with high school students who developed relationships with peers with handicaps. They found that the non-disabled students reported improved self-concept, reduced fear of human differences, increased tolerance of others, development of personal principles, and more relaxed and accepting friendships. Some difficulties they reported included social discomfort and discomfort in reaction to the physical characteristics of some students. Kamps et al. (1998) surveyed elementary school children who had been involved in social interventions for children with autism and found that peers were accepting and sometimes excited about participating in the activities with the children with autism. In addition, Pierce and Schreibman (1995) found that peer trainers reported that play sessions were challenging but also rewarding and educational. Thus, there is substantial evidence of the positive effects of these interventions for peers as well, including changes in negative expectancies regarding children with disabilities.
An individual's behavior is often affected by the expectations that his or her peers bring to the situation (Jones, 1990). Therefore, it is essential that future studies on socialization interventions for children with autism focus on the impact of peer expectancies. Though using peers, rather than adults, as models has been suggested as the most effective strategy for increasing the interactions of children with autism (Rogers, 2000), the importance of peer expectancies in their role as intervention agents has not been examined. Research has established the influence of teacher expectations on student behavior (e.g., Rosenthal, 1963), and this principle must not be forgotten when the role of teacher is being played by a peer.
Research has shown that without intervention, typically developing peers prefer to interact with each other (e.g., Myles et al., 1993). Therefore, specific strategies must be included in socialization interventions to change their expectancies regarding interactions with the children with autism. Based on the affect/effort theory, it is anticipated that a change in affect and effort should be noted following a change in expectancy (Rosenthal, 1989). The attending of children with autism to peer models would likely be fostered by an increase in peer effort to interact with them.
The Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (American Psychiatric Association, 1994) describes children with autism as having markedly impaired awareness of others and states that "individuals with this disorder may be oblivious to other children" (p. 66). Social learning theory suggests that one acquires new behavior by seeing the behavior modeled and receiving reinforcement, but this can be effective only if the child with autism pays attention to the model (Bandura, 1977). Yet, interventions rarely focus on measuring the attention paid by the children with autism to the intervention agent. The role of changes in peer expectancies to promote attention to the peer models is also often overlooked.
There is evidence that expectancies can be altered when the perceiver is made aware of them and is sufficiently motivated (Darley & Oleson, 1993). Therefore, interventions have the potential to promote peer effort through changing peer expectancies regarding children with autism. The present review suggests that there are several ways to do this, including arranging the situation or contingencies to promote optimal peer effort, teaching peers skills for initiating and reinforcing children with autism, and teaching the children with autism social skills. While all of these are plausible mechanisms for altering peer expectancies and promoting peer effort, the studies generally did not focus on this as a specific intervention goal. To promote optimal learning with peer-mediated strategies, peer expectancies should be more carefully studied.
There is evidence that social interventions involving peers hold promise for enhancing the socialization of children with autism. The research on peer-mediated interventions is growing, but there is still a great deal to be undertaken. Improved generalizability and social validity should be primary goals of future research. More comparative studies are needed to determine which interventions are most effective, and for which outcomes. Similarly, studies are needed to determine which components of the interventions result in the positive effects (Pollard, 1998). For example, details such as how many trained peers are necessary for increases in interaction and generalization should be more carefully considered (Mundschenk & Sasso, 1995). In addition, the impact of how peers are chosen should be systematically studied, with investigation of whether or not random assignment leads to increased generalizability. Characteristics of the children with autism, such as severity of impairment and age, should also be more carefully studied.
An overlooked but very important area of consideration is the role of peer expectancies. Altering peer expectations so that peers make greater effort to involve children with autism is one tactic for increasing attention. Poor attending to the environment is an undisputed area of deficit for children with autism, and change in that alone would be a positive step toward socialization. In addition, the importance of attention in learning (Bandura, 1977) must not be underestimated. Future studies should focus on both enhancing and measuring attention. The methods for promoting peer effort should be compared to determine which strategy most effectively draws the attention of the children with autism. More direct approaches for increasing attention should also be considered, such as providing both visual and verbal cues. Finally, learning should be measured in addition to performance (Bandura, 1977). Typically, studies on interventions to increase the socialization of children with autism measure changes in overt behaviors, such as the frequency or duration of social interactions, and the learning process is often overlooked. Pinpointing when and how learning best occurs will lead to greater refinement of intervention approaches.