Effects of Peer-Mediated Implementation of Visual Scripts in Middle School

Abstract

Although research has investigated the impact of peer-mediated interventions and visual scripts on social and communication skills in children with autism spectrum disorders, no studies to date have investigated peer-mediated implementation of scripts. This study investigated the effects of peer-implemented scripts on a middle school student with autism, intellectual impairments, and speech-language impairment via a multiple baseline single-case research design across behaviors. The target student demonstrated improvements in three communicative behaviors when implemented by a trained peer; however, behaviors did not generalize to use with an untrained typically developing peer.

Keywords

autism visual scripts communication peer-mediated intervention middle school adolescents

Many individuals with autism spectrum disorders (ASD) have difficulties with conversation skills such as initiation, eye contact, and reciprocal exchanges (Ganz, Kaylor, Bourgeois, & Hadden, 2008; Janzen, 2003). Thus, for these individuals, interventions that teach social skills and communication are necessary to improve their functioning within society (Chan et al., 2009; Harper, Symon, & Frea, 2008). As students get older, instruction tends to focus more on academic tasks rather than on social skills (Harper et al., 2008). Social skills instruction often takes place in contrived situations and settings and with the expectation that the student will generalize and use the skills in more natural settings (Gonzalez-Lopez & Kamps, 1997; Ladd & Asher, 1985). However, recent studies have investigated teaching social skills in the environment in which the skills will be used, such as on the playground or in play groups (Ganz & Flores, 2008; Harper et al., 2008). Scripts and peer-mediated interventions (PMI) are two strategies that have been used to improve social skills for individuals with ASD in natural settings (Chan et al., 2009; Goldstein, 2002).

Visual scripts have been used to promote social interactions for individuals with ASD (Ganz et al., 2008; Goldstein, 2002). Scripts have been used to teach a variety of communicative skills, such as requesting items or attention (Krantz & McClannahan, 1998), and conversation skills, such as making statements or asking questions of the communicative partner (Charlop-Christy & Kelso, 2003; Krantz & McClannahan, 1993). Scripts have been used with preschool (Goldstein & Cisar, 1992), elementary school (Ganz & Flores, 2008; Krantz & McClannahan, 1993), and middle school (Stevenson, Krantz, & McClannahan, 2000) children. Furthermore, scripts have been implemented with both readers (Ganz et al., 2008; Sarakoff, Taylor, & Poulson, 2001) and nonreaders (Krantz & McClannahan, 1998; Stevenson et al., 2000).

When implementing scripted communication interventions, the interventionist typically begins with a training session in which the individual with ASD is taught what the script says, via behavioral prompting techniques and prompt fading, until he or she is able to recite the script on seeing it. This is often followed by role-playing situations in which the scripts can be used. The scripts are then implemented during appropriate activities and prompted as needed to ensure proper use of the scripts during the activity. Krantz and McClannahan (1993) introduced a script-fading procedure in which the scripts were faded over time once the individual became independent with using the scripts. The scripts were faded from back to front; for instance, first, ending punctuation was removed, then sections of each phrase were removed from the end of the sentence until the paper was blank (Krantz & McClannahan, 1993, 1998). This procedure has been shown to improve both scripted and unscripted responses for individuals with ASD (Ganz et al., 2008; Krantz & McClannahan, 1993, 1998).

The National Professional Development Center (NPDC) on Autism Spectrum Disorder lists PMI as an evidence-based practice for students with ASD (Neitzel, Boyd, Odom, & Edmondson Pretzel, 2008). PMI has been defined as the use of peers as intervention agents to improve behaviors (Morrison, Kamps, Garcia, & Parker, 2001). PMI focuses on teaching both the typically developing peer and the individual with ASD how to interact with one another by prompting, reinforcing, and modeling appropriate skills (Haring & Breen, 1992). In support of recognition of PMI as an evidence-based practice, the NPDC found 11 studies that met the rigorous standards for high-quality research (Blew, Schwartz, & Luce, 1985; Garfinkle and Schwartz, 2002; Garrison-Harrell & Kamps, 1997; Goldstein, Kaczmarek, Pennington, & Shafer, 1992; Haring, & Breen, 1992; Kohler et al., 1995; Laushey & Heflin, 2000; Lee & Odom, 1996; Lee, Odom, & Loftin, 2007; Odom & Strain, 1986; Sainato, Goldstein, & Strain, 1992; Theimann & Goldstein, 2004). Furthermore, Chan et al. (2009) found 42 studies that utilized PMI as the primary intervention.

Schools and communities have an abundance of peers to use to help promote skill acquisition in individuals with ASD (Chan et al., 2009). The use of peers as interventionists may allow individuals with ASD to have more access to inclusive settings (Chan et al., 2009). Unfortunately, much of the research on PMI has focused on younger individuals, particularly on improving play skills and peers providing social reinforcement (Chan et al., 2009; Neitzel et al., 2008).

Many studies have used PMI to increase social interactions and improve social relationships for individuals with ASD (e.g., Goldstein et al., 1992; Gonzalez-Lopez & Kamps, 1997; Harper et al., 2008; Owen-DeSchryver, Carr, Calle, & Blakely-Smith, 2008; Shafer, Egel, & Neef, 1984; Theimann & Goldstein, 2001, 2004). Theimann and Goldstein (2001) combined social stories, written text cues (scripts), and video feedback with PMI to improve social interactions for an individual with ASD. The intervention package was effective in improving the student’s social functioning, but, due to the multicomponent nature of the package, it is unclear which strategy was most beneficial (Theimann & Goldstein, 2004). Therefore, Theimann and Goldstein (2004) isolated PMI and then added written text training prior to the second phase of the intervention. The teacher provided explicit instruction to both the peers and the individual with ASD to create Comic Strip Conversation© cues with phrases that could be used within the conversation. During the intervention sessions, the scripts were available to cue the discussion. The teacher prompted the use of the scripts when conversation lagged or students got off topic. PMI on its own was not as effective as the written cues in improving the student’s use of the targeted skills (Theimann & Goldstein, 2004).

To date, no studies have been conducted in which the typically developing peer prompted the scripted responses in the midst of a social activity without direct intervention from the teacher or researcher. In all previous script studies (e.g., Ganz et al., 2008; Krantz & McClannahan, 1993; Theimann & Goldstein, 2004), the interventionists were directly involved in training, prompting, and reinforcing the use of the scripts throughout the intervention sessions. The present study investigated the effectiveness of PMI and visual scripts on the use of communicative behaviors during leisure activities. The research questions were as follows: (a) Do peer-prompted script interventions result in increases in use of questions, praise, and requests for help in a middle school student with autism? (b) Is the use of the scripted phrases maintained as new skills are taught? (c) Does an individual with autism generalize the use of scripted phrases with an untrained peer? (d) Does an individual with autism continue using the scripted phrases as they are faded? and (e) Are the types of communicative functions learned and demonstrated by the student with autism at approximately the same level as used by typically developing peers?

Method

Participants and Setting

Participants were recruited through a local public school district, following university Institutional Review Board and school district research committee approval. A middle school teacher who worked with children with ASD nominated participants and sent home consent forms. The participants nominated by the teacher met the following selection criteria: (a) independent diagnosis of an ASD, (b) speech delay compared with same-age/grade peers, (c) enrollment in a public middle school, and (d) participation in school programming in which the student had minimum of a 30-min daily opportunity to interact with typically developing peers. One individual with autism returned her consent form. The teacher then sent consent forms to the parents of typically developing middle school peers who volunteered in the special education classroom. Once two typically developing peers returned their consent forms, the researchers then obtained assent from the participants.

Barbara was 15 years old at the beginning of the research study and was diagnosed with intellectual disability, autism, and speech impairment. According to her most recent Full and Individual Evaluation (FIE) given at age 13, Barbara used 1- to 2-word phrases spontaneously and could imitate 3- to 4-word phrases, which is considered to be within the “word combination” language phase (Tager-Flusberg et al., 2009). At the time of the study, she did not use an augmentative and alternative communication system. Barbara’s home language was Spanish, but all instruction at her school was given in English. In her most current FIE, the Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Renner, 1988) and Gilliam Autism Rating Scale–Second Edition (GARS-2; Gilliam, 1995) were used to determine that Barbara qualified for special education services as an individual with moderate to severe autism. She exhibited functional impairments in social interactions and communication, and displayed repetitive and stereotypic behaviors. Further details regarding Barbara’s assessments, including adaptive behavior, are provided in Table 1.

Table 1.

Participant Assessment Information

Assessment	Subscale—Score	Interpretation	Age at time of assessment
CARS	39.5	Severe autism	13
GARS-2	Stereotyped Behavior—10 Communication—8 Social Interaction—9 Autism Quotient—94	High probability of autism	13
DP-II	Physical—42 Self Help—46 Social—32 Academic—24 Communication—28 IQ Equivalence—29	Moderate cognitive impairment	6
Vineland-II	Communication—57 Daily Living—58 Socialization—60 Motor Skills—65 Adaptive Behavior Comp.—56	Moderate cognitive impairment	6

Note: CARS = The Childhood Autism Rating Scale (Schopler, Reichler, & Renner, 1988); GARS-2 = Gilliam Autism Rating Scale (Gilliam, 1995); DP-II = Developmental Profile–II (Alpern, Boll, & Shearer, 2000); Vineland-II = Vineland Adaptive Behavior Scales–Second Edition (Sparrow, Cicchetti, & Balla, 2005).

The parents of two typically developing middle school peers consented, and the students assented to participate in the study. Debby and Emilia attended Barbara’s special education class as an elective. Prior to this study, they received introductory instruction from Barbara’s teacher on how to interact with individuals with autism and had been provided with basic information regarding what autism is and how it affects the individual’s ability to learn new skills. The peers were randomly assigned to be either the trained peer or the untrained peer. Debby was assigned to be the trained peer, whereas Emilia served as the untrained peer. Debby was asked not to provide Emilia with information about the study, other than that she was working on craft activities with Barbara.

The intervention took place at a table in a cooking lab across the hall from Barbara’s classroom. This setting allowed for minimal disruption, was a familiar environment for Barbara, and allowed video recording for data collection; a quiet room was required because Barbara spoke with low volume and was difficult to hear on video when in the classroom. During each session, crafting materials were provided at the table and the typically developing peer sat next to Barbara with the materials available for use directly in front of them on the table.

Materials

The authors asked Barbara’s teacher to suggest two activities that were age appropriate and that she thought the students would enjoy. The teacher reported that Barbara enjoyed craft activities such as coloring and stamping. Card making and jewelry making were chosen as the two activities for the students to engage in. For the card-making activity, a variety of colored cardstock paper, stamps, ink pads, stickers, and markers were available for use. For the jewelry-making activity, the participants had access to a variety of beads as well as elastic clear thread.

The researchers developed scripts for each communicative response, questioning, praise, comments, and requests, with line drawings for each phrase. The scripts were printed on regular printer paper. Each script was limited to 2- to 3-word phrases based on Barbara’s communication assessment provided in the FIE. A 1-inch picture was included with the written script because Barbara did not demonstrate reading skills at the whole-word level. See the appendix for a sample of each script.

Response Definitions and Measurement

Data were collected on four communicative responses: question, praise, comments, and requests. The appendix provides a condensed version of the scripts used for each communicative response and examples of verbal responses. Questions were defined as questions posed to the other person either in context with the activity or otherwise. Praise was defined as saying something complimentary about the other participant’s work, appearance, or any other topic. Comments were defined as any verbal response that was not praise or questioning; no script was used for commenting. After collecting baseline data on commenting during the praise and questioning phases, it was decided that Barbara used an appropriate level of commenting but did not request help even when she clearly needed assistance. Therefore, requesting help was added as the final communicative response. Requesting help was defined as any verbal attempt to get the other person to give aid or help with any situation.

Intervention sessions were 3 min in length. Data were collected on all communicative behaviors during every data probe. Data were collected using partial interval recording with 15-s intervals. If at any point within an interval the behavior was displayed, the researcher circled the code for that behavior. The participant was not required to use the exact wording written on a script. If none of the behaviors were used during that interval, the researcher marked the code for “none.” Data were reported as percentage of intervals each behavior was demonstrated during the intervention session. Two intervention sessions were held each day with a brief break between the sessions to provide instruction on the scripts. Data were collected on all behaviors during every session for both participants.

Throughout the study, data were collected on typical peer interactions during the same activity to serve as a basis for determining mastery and provide a comparison to help determine if Barbara’s behavior was approximating typical peer levels. In these sessions, Debby and Emilia used the same materials in the same setting as Barbara. Data were collected using the same data collection sheet and interval recording process.

Experimental Design

A concurrent multiple baseline design across communicative responses was used to determine the effect of PMI combined with scripts as an intervention package. Generalization data probes were imbedded into each phase of the study. As the intervention moved to the next phase, the prior scripts were no longer prompted, and therefore the prior skills were placed in a maintenance phase. Once all of the scripts had been implemented and the participant demonstrated treatment effects for the final communicative response, all scripts were available and were prompted during the next phase. Finally, scripts were faded in the final phase.

Procedures

Baseline

During baseline condition, all materials were present for the activity, but no scripts were provided. Prior to baseline sessions, Debby was told to interact with Barbara as usual, and no further instruction was provided for Barbara or Debby. During baseline, no generalization data were collected because no skills had been mastered. During the following phases, generalization data were taken as each skill was mastered and no longer prompted.

Intervention

Before each intervention data session, a researcher briefly taught or reviewed with Debby how to work with Barbara with the scripts and appropriate prompting strategies. Debby was provided with a short training script with bulleted points to help remind her to use the strategies taught in the training sessions while working with Barbara. It was left up to Debby to decide if she wanted to keep her training points on the table and visible during the sessions or put them away. Most of the time, Debby put them away during intervention sessions. Following these training sessions, Debby participated in a brief training session with Barbara to practice the scripts. The researchers modeled how to provide verbal and gestural prompts to help Barbara learn what to say for each script. Debby was then given a chance to work with Barbara with feedback from the researchers. The students then began a craft activity (intervention data session), during which Debby prompted Barbara at pertinent times within the activity to use the scripted phrase by pointing to the phrase and then, if Barbara did not recite one of the script phrases, verbally modeled the script.

During the 1st week of intervention sessions, Barbara needed verbal models. The researchers began intensive training sessions prior to each activity session. The intensive training sessions began with the interventionist holding up the script, pointing to the words, and verbally modeling the scripted phrase. The interventionist then prompted Barbara by saying “your turn.” If Barbara correctly repeated the phrase she was given praise, and the interventionist moved to the next script phrase. After the first two to three correct responses with a verbal model, the verbal model was delayed 3 to 5 s to give Barbara a chance to respond without the prompt. The training session continued until Barbara correctly said each scripted phrase without a verbal model. After the training session, a craft session was held in which Debby was instructed to use gestural prompts and no verbal models. During the crafting sessions (intervention data sessions), the scripts were left on the table between Barbara and the peer so that Debby could prompt the scripts without reaching across Barbara. One generalization probe was collected per week throughout intervention.

Each phase of intervention lasted 2 weeks. In the 1st week, verbal models were used, whereas in the 2nd week, only gestural prompts were provided by the peer during intervention data sessions. Once each phase was complete, the previously taught scripts moved into maintenance and were no longer prompted. For praise and questions, two scripts were taught to provide a variety of response options. In the final phase, only one script was taught to request help.

For Phase 3 of intervention, Barbara was introduced to a single script with the prompt to request help. This phase differed slightly from the previous instructional phases because Debby was instructed to “set up” a situation to ensure that Barbara had opportunities to request help. During baseline data collection for requesting help, there was at least one opportunity per session for Barbara to ask for help. During intervention, Debby was instructed to ensure at least one and preferably two or more opportunities to request help. Debby was allowed to manipulate the materials to allow for more opportunities to communicate, such as leaving lids on the bead containers to allow Barbara an opportunity to request help to open the containers. Debby was instructed to wait for Barbara to briefly attempt the task unsuccessfully before prompting the script for help to ensure that Barbara did need help with the task.

Once requesting help had been prompted for 2 weeks, the final phase of intervention included prompting all scripts during the intervention session. Debby was told to prompt the scripts as it felt natural within the training session. As with all other phases, Debby was allowed to use verbal models for the 1st week, but the 2nd week only gestural prompts were used. Intensive training sessions for this phase, which occurred prior to data collection sessions, included prompting for every script until Barbara no longer needed a verbal model. During the craft sessions, data were collected as usual.

Script fading

The final phase involved fading the words and pictures from the scripts. This was completed in a three-step process: First, one fourth of the image and words were removed from the end of each script, then one half of the image and words were removed, and finally the script cards were blank.

Interobserver Agreement (IOA)

IOA was collected by two researchers for 19% of Barbara and Debby’s data. Due to an oversight in scheduling video recording, Emilia’s (generalization) data were only observed by two observers one time. IOA was calculated by determining the number of agreements and disagreements interval-by-interval and dividing the number of agreements by agreements plus disagreements and multiplying times 100 (Kazdin, 1982). IOA averaged 95% (range = 83%-100%) for Barbara’s data and 89% (range = 81%-100%) for Debby’s data; IOA was 97% for the single generalization data point collected for Barbara and 100% for Emilia.

Analyses

Data were graphed and analyzed via visual analysis to inspect for data point level change, trend, and intercept gap to determine if there was a functional relationship between the intervention and the participant’s use of communicative skills (Kazdin, 1982). Visual analysis was supplemented with an effect size measure, the Nonoverlap of All Pairs (NAP; Parker & Vannest, 2009). Area under the curve (AUC) outputs from a receiver operator characteristic (ROC) curve analysis, which is a well-established, statistically robust measure (D’Agostino, Campbell, & Greenhouse, 2006), is equivalent to NAP. NAP is a nonparametric effect size that calculates nonoverlap, or improvement in data points, between phases (Parker, Vannest, & Davis, in press). NAP scores range from 0 to 1.0. A score of .5 indicates no difference between phases, above .5 indicates improvement from baseline to intervention, and below .5 indicates decline in performance (Parker & Vannest, 2009).

NAP has only recently begun to be used to interpret single-case research (Parker & Vannest, 2009). However, use of effect sizes in single-case research has increased and has begun to gain acceptance as a supplement to visual analysis in current standards for evaluating the body of single-case intervention research (Horner et al., 2005; Kratochwill et al., 2010). In this study, NAP scores and 95% confidence intervals (CIs) were calculated between the baseline phase and intervention phase for each of the three communicative behaviors for Barbara.

Results

Results for Barbara’s use of communicative behaviors are found in Figure 1, and the typically developing peers’ use of communicative behaviors is illustrated in Figure 2. The x axes in both figures represent sessions, and the y axes represent percentage of intervals during which the participants demonstrated the behaviors.

Figure 1.

Barbara’s use of communicative behaviors

Figure 2.

Typically developing peers’ use of communicative behaviors

Barbara

Barbara used few questions during baseline (M = 5%; range = 0%-8%). Her performance quickly increased during intervention for questions, although her performance was variable (M = 33%; range = 17%-50%). When intervention ceased for questions, she continued to use some questions initially, but there was a rapid decreasing trend, followed by a long span of the use of few questions (M = 4%; range = 0%-42%). When prompts for all communicative acts resumed, Barbara’s use of questions quickly increased, though still variable and at lower levels than the initial intervention phase (M = 17%; range = 8%-33%). As the scripts were faded, Barbara continued to use questions at similar levels to those exhibited during the reintroduction of all scripts; these levels were slightly higher than baseline, although the data overlap considerably (M = 9%; range = 0%-17%). During maintenance, Barbara used questions during 8% of the intervals for both data points, which was nearly at the same level as her performance during the fading phase. Barbara’s generalization data (with Emilia) demonstrate low levels of question use in all phases, at levels that were lower than during Barbara’s performance with the peer interventionist (Debby).

During baseline, Barbara used no praise statements. These increased rapidly, with an increasing trend during script intervention for praise (M = 21%; range = 8%-33%). When scripts for praise were withdrawn, her use dropped, although not as low as during baseline (M = 2%; range = 0%-8%). That is, she used no praise during some sessions and a small number of praise statements during others. When all scripts were reintroduced, she used praise statements again in approximately half of the sessions, with an increasing trend and levels near those during the initial intervention phase (M = 9%; range = 0%-25%). She continued using the praise statements throughout the fading procedures (M = 15%; range = 0%-25%) and used them during 0% of intervals during one maintenance session and 25% of the other. Barbara did not use any praise statements during generalization sessions with Emilia, during any phase.

During baseline, Barbara did not request help. She quickly responded to intervention with a higher, although variable, level of requests for help (M = 18%; range = 8%-25%). Scripts for requests for help were not immediately withdrawn, but Barbara maintained use of help requests while all of the scripts were being prompted, although her performance was somewhat variable (M = 12%; range = 0%-25%). Her requests for help decreased slightly during the fading procedures; however, she continued to request help during nearly all sessions, and her performance was less variable (M = 9%; range = 0%-25%); she requested help during 17% of intervals during one maintenance session and 8% of the other. She did not make any requests for help during any generalization sessions with Emilia.

Overall, a strong functional relationship was demonstrated between the peer-mediated script intervention and Barbara’s use of three types of communicative acts, based on visual analysis of the data during intervention with Debby, the peer who received instruction and who provided script-prompting and fading procedures. NAP effect size analyses corroborated the visual analyses, demonstrating statistically significant, strong effects for all three communicative behaviors: questions (NAP = 1.00; 95% CI = [0.91, 1.00]; p = .0005), praise (NAP = 1.00; 95% CI = [0.92, 1.00]; p = .0000), and requests (NAP = 1.00; 95% CI = [0.89, 1.00]; p = .0000).

Barbara Compared With Peers

Debby and Emilia used the three communicative skills at similar levels to one another throughout data collection. Compared with her typically developing peers, Barbara used fewer questions during all phases; however, her use during the initial intervention phase with questions approached levels that were typical of her peers. Barbara’s use of praise was at similar levels to her typically developing peers throughout the study; however, both of the peers used praise at least occasionally during baseline, whereas Barbara did not until intervention commenced. Barbara’s peers nearly never requested help; therefore, beyond baseline, Barbara requested help more frequently than her peers.

Discussion

The present study examined the use of a scripted PMI to increase functional and social communicative behaviors (i.e., questioning, praising, and requesting help) in a middle school student with moderate to severe autism. Additional data were collected to investigate generalization of the taught communication skills with an untrained peer. Furthermore, data on the interactions initiated by two typically developing students were collected to allow for comparison of Barbara’s interactions to those of two typically developing peers.

Barbara used each communicative behavior very infrequently, if at all, during baseline data collection. Her use of the communicative behaviors increased significantly while she was prompted with the script by her peer activity partner. That is, she used few to none of the behaviors during baseline, but increased to low levels of each behavior throughout each data probe, during intervention. Low levels were expected because Barbara used nearly none prior to intervention. One would not expect a student to praise another student’s work more than once or twice during a 3-min session. Furthermore, levels of these behaviors were similar to the levels of the behaviors used by Barbara’s peers (see Figure 2). When scripts were withdrawn, Barbara’s use of the communicative behaviors dropped but usually remained slightly above baseline levels. When the use of all scripts was reintroduced, Barbara’s use of the communicative behaviors generally increased again and was maintained through fading procedures. Although her use of the communicative behaviors during maintenance was again variable, Barbara’s maintenance levels of usage were generally above her baseline levels. Barbara seldom used communicative behaviors during generalization sessions with an untrained peer. Her level of use of the targeted communicative behaviors during intervention sessions was roughly equivalent to that of two typically developing peers.

The current study supplements and extends previous research demonstrating the effectiveness of implementation of visual scripts with children and youth with autism and typically developing peers (Ganz & Flores, 2008) and with middle school–aged students (Stevenson et al., 2000). Furthermore, this study provides modest support for the use of PMI and visual scripts to increase communicative behaviors in an adolescent with autism. Barbara’s use of the targeted communicative behaviors increased significantly once the script prompts were introduced and decreased when the scripts were withdrawn, suggesting a direct relationship between peer prompting with the scripts and communicative behaviors. Furthermore, Barbara’s continued use of the communicative behaviors at levels slightly above baseline after the scripts were withdrawn suggests that the use of scripts with PMI may be an effective tool for teaching and reinforcing functional and social communicative behavior. This is also supported by the fact that Barbara’s use of communicative behaviors, although variable, generally remained above baseline during script fading and maintenance.

Social Validity and Generalization

One of the chief advantages of PMI is that they can easily fit within natural, age-appropriate settings and activities and therefore may encourage more natural social interactions. However, due to the somewhat contrived nature of some research studies, the extent to which the setting and peer interactions are naturalistic can vary. For instance, the typically developing peers, Debby and Emilia, in this study were aware that they were taking part in a research study, and Debby knew that she was teaching Barbara. Perhaps accordingly, both peers started with very “teacher”-like language in which they were instructing Barbara how to complete the task. As the study progressed, Debby’s interactions became more interactive and friendly whereas Emilia remained more distant and teacher-like, possibly due to her relatively limited interaction time with Barbara.

There are several possible reasons Barbara did not generalize her use of newly learned communication behavior to use with Emilia. First, the lack of natural rapport may help explain the poor generalization of communicative behaviors during Barbara’s sessions with Emilia. Barbara may have felt anxious or unsure around this new peer and could have seen her as another “teacher” rather than a peer, creating an imperfect generalization scenario. Also, given that autism is often characterized by difficulty with change and anxiety or avoidance regarding socializing with unfamiliar people (American Psychiatric Association, 2000), it is possible that Emilia’s status as a relatively unknown person may have created additional anxiety for Barbara, leading to the apparent lack of generalization.

Generalization may have been improved had a number of interventions been implemented. First, the visuals that were implemented with Debby could have been present on the table during generalization probes with Emilia. Even without peer prompting, the visual supports may have served as a means for Barbara to cue herself to use the learned phrases. Furthermore, more targeted training could have been provided to Emilia or both peers regarding characteristics of autism and means of interacting effectively. Finally, a brief training session could have been provided for Emilia, either by the investigators or with Debby’s assistance, as might occur during natural events in integrated settings; during the study, Debby was instructed not to discuss the procedures with Emilia.

Choice of Activity and Communicative Behavior

Questions regarding how to select an appropriate activity and communicative behavior targets are also considerations when conducting PMI research. In this study, the researchers made a concentrated effort to choose a feasible, age-appropriate, and developmentally appropriate activity that both Barbara and the peer mentors would reasonably engage in within natural contexts. However, the level of difficulty of the activity may have been mismatched, given that Barbara had to request help whereas neither the two typically developing peers required help, making it difficult to compare Barbara’s help-requesting behavior with that of her peers without disabilities. In addition, it is possible that the interactions of the typically developing peers may have been influenced by the fact that they knew their behavior was being recorded for research purposes. Despite these limitations, however, it remains likely that the activity and circumstances (i.e., same age peer interaction) in this study more closely approximate natural interactions than would traditional research scenarios that involve direct social interaction with the investigators. This could possibly lead to improved maintenance of the communicative behaviors postintervention.

Methodological Limitations

Although extensive efforts were made to use a rigorous single-case methodology, this study would have been strengthened by the collection of more data on IOA, particularly for generalization probes. The existing IOA data suggest strong agreement between raters, and more data would lend additional support to the assumption that observers overwhelmingly agreed on the presence or absence of communicative behavior during each recording interval. Furthermore, NAP results should be viewed with caution in consideration of the early stage of use and acceptance of effect size measures in single-case research (Horner et al., 2005; Kratochwill et al., 2010).

Implications and Directions for Research and Practice

This study demonstrates that PMI can be feasibly used with middle school–age students, expanding the scope of PMI research and practice beyond preschool age students, who have traditionally been the focus of PMI interventions (Chan et al., 2009; Neitzel et al., 2008). In addition, this study demonstrates that PMI can be implemented effectively by trained middle school–age peers without direct teacher or researcher involvement. This has the potential to reduce the staff burden of the intervention while providing for more naturalistic interaction. Although there was an “up front” time cost in training the trained peer to use the scripts, the subsequent burden on researchers and staff was limited to setting up the activity area and recording data, versus intensive, direct involvement in each phase and intervention session. This may increase the feasibility of PMI interventions in this age group. Future research should examine the necessary time required to train typically developing peers to use scripts with peers with ASD and other disabilities and examine ways to increase the efficiency of peer training. Future studies should also include collection of distal outcomes data, such as on quality and quantity of interactions with peers outside of study procedures.

Due to the inherent limitations on the conclusions that can be drawn from a single-case study, it is important that researchers continue to examine the efficacy and use of PMI with middle school students. In particular, researchers should investigate ways to increase generalizability of communication to interactions with untrained peers and develop guidelines for selecting activities that are likely to prompt similar communicative behaviors (e.g., help seeking) from typically developing peers and students with developmental disabilities. The use of PMI with middle school students is a relatively nascent area of research but a promising one.

Footnotes

Appendix

Acknowledgements

The materials for this study were provided by the Center on Disability and Development at Texas A&M University, a federally designated University Center for Excellence in Developmental Disabilities.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: Graduate student assistance was funded by the Department of Educational Psychology at Texas A&M University.

Bios

Jennifer B. Ganz, PhD, BCBA-D, is an associate professor of special education at Texas A&M University. She has published numerous research articles involving interventions to improve communication and social skills in children with autism spectrum disorders and intellectual/developmental disabilities.

Amy K. Heath is an autism specialist in Bryan Independent School District in Bryan, TX, and is a doctoral candidate in special education at Texas A&M University.

Emily M. Lund is a graduate student in school psychology at Texas A&M University.

Siglia P. H. Camargo is a doctoral student in special education at Texas A&M University.

Mandy J. Rispoli, PhD, BCBA-D, is an assistant professor of special education at Texas A&M University. She has published extensively, reporting research involving function-based interventions to improve challenging behavior in children with autism spectrum disorders and intellectual/developmental disabilities.

Margot Boles is a graduate student in special education at Texas A&M University.

Lauren Plaisance is a graduate student in behavior analysis at the University of Houston, Clear Lake.

References

Alpern

Boll

Shearer

(2000). Developmental profile II manual. Los Angeles, CA: Western Psychological Services.

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

Blew

P. A.

Schwartz

I. S.

Luce

S. C.

(1985). Teaching functional community-based skills to autistic children using nonhandicapped peer tutors. Journal of Applied Behavior Analysis, 18, 337-342.

Chan

Lang

Rispoli

O’Reilly

Sigafoos

Cole

(2009). Use of peer-mediated interventions in the treatment of autism spectrum disorders: A systematic review. Research in Autism Spectrum Disorders, 3, 876-889.

Charlop-Christy

M. H.

Kelso

S. E.

(2003). Teaching children with autism conversational speech using a cue card/written script program. Education and Treatment of Children, 26, 108-127.

D’Agostino

R. B.

Campbell

M. J.

Greenhouse

J. B.

(2006). Editorial: The Mann-Whitney statistic: Continuous use and discovery. Statistics in Medicine, 25, 541-542.

Ganz

J. B.

Flores

M. M.

(2008). Effects of the use of visual strategies in play groups for children with autism spectrum disorders and their peers. Journal of Autism and Developmental Disorders, 38, 926-940.

Ganz

J. B.

Kaylor

Bourgeois

Hadden

(2008). The impact of social scripts and visual cues on verbal communication in three children with autism spectrum disorder. Focus on Autism and Other Developmental Disabilities, 23, 79-94.

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, 26-38.

10.

Garrison-Harrell

Kamps

Kravits

(1997). The effects of peer networks on social-communicative behaviors for students with autism. Focus on Autism and Other Developmental Disabilities, 12, 241-256.

11.

Gilliam

J. E.

(1995). Gilliam Autism Rating Scale. TX: Pro-ed.

12.

Goldstein

Cisar

C. L.

(1992). Promoting Interaction During Sociodramatic Play: Teaching Scripts To Typical Preschoolers And Classmates With Disabilities. Journal Of Applied Behavior Analysis, 25, 265-280.

13.

Goldstein

(2002). Communication interventions for children with autism: A review of treatment efficacy. Journal of Autism and Developmental Disorders, 17, 373-396.

14.

Goldstein

Kaczmarek

Pennington

Shafer

(1992). Peer-mediated intervention: Attending to, commenting on, and the behavior of preschoolers with autism. Journal of Applied Behavior Analysis, 25, 289-305.

15.

Gonzalez-Lopez

Kamps

(1997). Social skills training to increase social interactions between children with autism and their typical peers. Focus on Autism and Other Developmental Disabilities, 12, 2-14.

16.

Haring

Breen

(1992). A peer-mediated social network intervention to enhance the social integration of persons with moderate and severe disabilities. Journal of Applied Analysis, 25, 319-333.

17.

Harper

C. B.

Symon

J. B. G.

Frea

W. D.

(2008). Recess is time-in: Using peers to improve social skills of children with autism. Journal of Autism and Developmental Disorders, 38, 815-826.

18.

Horner

R. H.

Carr

E. G.

Halle

McGee

Odom

Wolery

(2005). The use of single-case research to identify evidence-based practice in special education. Exceptional Children, 71, 165-179.

19.

Janzen

J. E.

(2003). Understanding the nature of autism: A guide to the autism spectrums disorders (2nd ed.). San Antonio, TX: Therapy Skill Builders.

20.

Kazdin

A. E.

(1982). Single-case research designs: Methods for clinical and applied settings. New York, NY: Oxford University Press.

21.

Kohler

F. W.

Strain

P. S.

Hoyson

Davis

Donina

W. M.

Rapp

(1995). Using a group-oriented contingency to increase social interactions between children with autism and their peers: A preliminary analysis of corollary supportive behavior. Behavior Modification, 19, 10-32.

22.

Krantz

P. J.

McClannahan

L. E.

(1993). Teaching children with autism to initiate to peers: Effects of a script-fading procedure. Journal of Applied Behavior Analysis, 26, 121-132.

23.

Krantz

P. J.

McClannahan

L. E.

(1998). Social interaction skills for children with autism: A script-fading procedure for beginning readers. Journal of Applied Behavior Analysis, 31, 191-202.

24.

Kratochwill

T. R.

Hitchcock

Horner

R. H.

Levin

J. R.

Odom

S. L.

Rindskopf

D. M.

Shadish

W. R.

(2010). Single-case designs technical documentation. Retrieved from http://ies.ed.gov/ncee/wwc/pdf/wwc_scd.pdf

25.

Ladd

G. W.

Asher

S. R.

(1985). Social skills training and children’s peer relations. In L’Abate

L. L.

Milan

M. A.

(Eds.), Handbook of social skills training and research (pp. 219-244). New York, NY: John Wiley.

26.

Laushey

K. M.

Heflin

L. J.

(2000). Enhancing social skills of kindergarten children with autism through the training of multiple peer tutors. Journal of Autism and Developmental Disorders, 30, 183-193.

27.

Lee

Odom

S. L.

(1996). The relationship between stereotypic behavior and peer social interactions for children with severe disabilities. Journal of the Association for Persons with Severe Handicaps, 21, 88-95.

28.

Lee

Odom

S. L.

Loftin

(2007). Social engagement with peers and stereotypic behavior of children with autism. Journal of Positive Behavior Interventions, 9, 67-79.

29.

Morrison

Kamps

Garcia

Parker

(2001). Peer mediated and monitoring strategies to improve initiations and social skills for students with autism. Journal of Positive Behavior Interventions, 3, 237-250.

30.

Neitzel

Boyd

Odom

S. L.

Edmondson Pretzel

(2008). Peer-mediated instruction and intervention for children and youth with autism spectrum disorders: Online training module. Chapel Hill, NC: National Professional Development Center on Autism Spectrum Disorders, FPG Child Development Institute, UNC-Chapel Hill. Available from www.autisminternetmodules.org.

31.

Odom

S. L.

Strain

P. S.

(1986). A comparison of peer-initiation and teacher-antecedent interventions for promoting reciprocal social interactions of autistic preschoolers. Journal of Applied Behavior Analysis, 19, 58-72.

32.

Owen-DeSchryver

J. S.

Carr

E. G.

Calle

S. I.

Blakely-Smith

(2008). Promoting social interactions between students with autism spectrum disorders and their peers in inclusive school settings. Focus on Autism and Other Developmental Disabilities, 23, 15-28.

33.

Parker

R. I.

Vannest

K. J.

(2009). An improved effect size for single case research: Non-Overlap of All Pairs (NAP). Behavior Therapy, 40, 357-367. doi:10.1016/j.beth.2008.10.00610.1016/j.beth.2008.10.006

34.

Parker

R. I.

Vannest

K. J.

Davis

J. L.

(in press). Reliability for multi-category rating scales: A Monte-Carlo simulation. Journal of School Psychology.

35.

Sainato

D. M.

Goldstein

Strain

P. S.

(1992). Effects of self-evaluation on preschool children’s use of social interaction strategies with their classmates with autism. Journal of Applied Behavior Analysis, 25, 127-141.

36.

Sarakoff

R. A.

Taylor

B. A.

Poulson

C. L.

(2001). Teaching children with autism to engage in conversational exchanges: Script fading with embedded textual stimuli. Journal of Applied Behavior Analysis, 34, 81-84.

37.

Schopler

Reichler

Renner

(1988). The Childhood Autism Rating Scale (CARS). Los Angeles: Western Psychological Services.

38.

Shafer

M. S.

Egel

A. L.

Neef

N. A.

(1984). Training mildly handicapped peers to facilitate changes in the social interaction skills of autistic children. Journal of Applied Behavior Analysis, 17, 461-476.

39.

Sparrow

Cicchetti

Balla

(2005). Vineland Adaptive Behavior Scales Second Edition (Vineland-II). Circle Pines, MN: AGS.

40.

Stevenson

C. L.

Krantz

P. J.

McClannahan

L. E.

(2000). Social interaction skills for children with autism: A script-fading procedure for non-readers. Behavioral Interventions, 15, 1-20.

41.

Tager-Flusberg

Rogers

Cooper

Landa

Lord

Paul

Yoder

(2009). Defining spoken language benchmarks and selecting measures of expressive language development for young children with autism spectrum disorders. Journal of Speech, Language, and Hearing Research, 52, 643-652.

42.

Theimann

Goldstein

(2001). Social stories, written text cues, and video feedback: Effects on social communication of children with autism. Journal of Applied Behavior Analysis, 34, 425-446.

43.

Theimann

Goldstein

(2004). Effects of peer training and written text cueing on social communication of school-age children with pervasive developmental disorder. Journal of Speech, Language, and Hearing Research, 47, 126-144.