Abstract
Although research suggests peer support arrangements can be an effective practice for improving social outcomes for students with severe disabilities, additional efforts are needed to refine training and implementation approaches to increase the replicability and sustainability of this intervention. We tested a promising teacher-delivered training model using a multiple-probe-across-participants design. Four paraprofessionals received training to facilitate peer support arrangements for four middle school students with severe disabilities in general education classrooms. Teacher-delivered training enabled paraprofessionals to facilitate peer support arrangements that increased support from peers and interactions for students with disabilities. In addition, three of four students with disabilities demonstrated growth on individualized goals. We discuss implications for research and practice, highlighting approaches for improved implementation of peer support arrangements.
Keywords
Students with severe disabilities (i.e., students with intellectual disability and/or autism who qualify for their state’s alternate assessment) are increasingly spending time in general education classrooms (U.S. Department of Education, 2014). Among the desired outcomes of inclusion are increased social interactions and shared learning opportunities with peers who do not have disabilities (Jackson, Ryndak, & Wehmeyer, 2008–2009). Because the importance of peer relationships becomes more prominent in adolescence, efforts to foster social connections become particularly important for middle school students. Unfortunately, adolescents with severe disabilities interact infrequently with their peers and struggle to make social connections. Descriptive studies show that these students rarely interact with their peers, even when present in the same classroom (Carter, Sisco, Brown, Brickham, & Al-Khabbaz, 2008).
Several factors may contribute to this lack of interaction. First, students with severe disabilities are often seated away from their peers, decreasing the likelihood that interaction will occur naturally (Feldman, Carter, Asmus, & Brock, 2015). Second, peers are sometimes uncertain how to interact with students with severe disabilities, especially if they lack prior experience spending time with individuals who have complex communication and support needs. Third, students with severe disabilities are often provided individual adult support in general education classrooms. The constant presence of an adult can unintentionally deter peer interactions, particularly in adolescence (Carter et al., 2008; Giangreco, 2010). These contextual and peer factors may ultimately be more of an impediment to peer interaction than social- and communication-related skill deficits (Cutts & Sigafoos, 2001).
Peer support arrangements provide a means to directly address these barriers by increasing proximity to peers, boosting peer skills and confidence, and diminishing the barrier of individualized adult support (Carter, Moss, Hoffman, Chung, & Sisco, 2011). These interventions involve four key steps, including (a) designing an individualized peer support plan, (b) recruiting peers from within the same classroom, (c) orienting peers to their new roles, and (d) providing ongoing facilitation of the peer support arrangement. An adult facilitator who provides support to the student with a disability—typically a paraprofessional—works with a special educator to develop a written, individualized peer support plan. However, it is the licensed teacher who takes primary responsibility for developing the plan, seeking input from the paraprofessional who often is more familiar with typical routines and activities in the general education classroom. Peer support plans involve identifying appropriate ways for peers to provide support during different instructional activities or formats (e.g., large-group lectures, small-group activities, and independent seat work). Support behaviors target social interaction (e.g., greeting the student before class and engaging him in conversation about a common interest, introducing the student to another friend) and/or academic engagement (e.g., prompting the student through a modified academic task, identifying a role with which the student can be successful during a small-group activity). In addition, the written plan outlines how the facilitator will support peers in their roles. Next, the facilitator works with the general education teacher to identify peers who could provide support. The facilitator might individually approach peers with a positive history of interaction with the student, ask the student who he or she would like to work with, or extend an open invitation to the class for interested peers. The facilitator meets with interested peers to orient them to their new roles and share the strategies outlined in the peer support plan. Finally, the facilitator prompts and reinforces appropriate support behaviors used by peers in the classroom. In addition, the facilitator might provide additional information to the peers that helps them better support the student.
Peer support arrangements are an effective strategy to increase interactions between students with severe disabilities and their peers (Carter, Sisco, Chung, & Stanton-Chapman, 2010). However, two critical questions remain about how this strategy might be scaled up in schools. First, for this strategy to be replicable, the implementation behaviors associated with improved student outcomes must be well defined and measured. These behaviors include paraprofessional facilitation (i.e., the frequency and nature of how they promote peer support) and peer support behaviors (i.e., the frequency and nature of how they support students with severe disabilities). Second, for this strategy to be sustainable, someone other than a member of an external research team must take responsibility for training and supporting paraprofessional facilitators. In most efficacy studies, members of an external research team implement parts of the intervention themselves (e.g., Carter et al., 2011) or provided extensive support to school-based practitioners (e.g., Shukla, Kennedy, & Cushing, 1999). Identifying someone to fill this role is critical to ensure the success of peer support arrangements. These challenges are best solved through implementation science—the study of effective and efficient methods to put promising approaches into everyday practice (Cook & Odom, 2013).
Brock and Carter (2015) evaluated a model for replicable and sustainable implementation of peer support arrangements. In their pilot study, special educators delivered a training package to paraprofessionals that emphasized two promising training strategies: modeling and performance feedback (Brock & Carter, 2013). Teachers used a combination of live and video modeling to teach paraprofessionals to implement peer support arrangements. After paraprofessionals set up the intervention and began facilitating, teachers observed in the general education classroom and delivered performance feedback. During the performance feedback session, the teacher highlighted both the positive aspects of implementation she observed, as well as specific opportunities for improvement. This combination of strategies enabled four paraprofessionals to implement peer support arrangements for students with severe disabilities. Findings suggested low but consistent rates of facilitation are sufficient for students with severe disabilities to experience improved outcomes.
Although this pilot study advanced understanding of implementation behavior at the paraprofessional level, it did not closely examine a second level of implementation—peer behavior. Peer implementation behaviors have received fairly limited attention in prior studies examining the efficacy of peer support arrangements. To date, no study has reported rates of peer prompting, reinforcement, or use of specific strategies. The nature of how peers are actually providing support throughout the class period remains unclear, as does its connection to social-related outcomes. To fully understand and replicate implementation of peer support arrangements, it is critical to better understand the impact of all implementation behaviors: how training affects paraprofessional behavior, how paraprofessional behavior affects peer behavior, and how peer behavior affects outcomes for students with disabilities. Currently, no published study has measured or examined the relationships between all these variables.
The present study directly addresses this problem by replicating teacher-delivered training for paraprofessionals, and measuring (a) specific paraprofessional facilitation behaviors; (b) the degree to which peers use prompting, reinforcement, and individualized intervention strategies to support their classmates with disabilities; and (c) the degree to which paraprofessional and peer behaviors generalize to a new classroom setting.
Method
Students With Disabilities, Paraprofessionals, and Special Educators
After receiving Institutional Review Board and district approval, we recruited four triads that each included one middle school student with severe disabilities, one paraprofessional, and one supervising special education teacher. Students included in the study had to be (a) receiving special education services under the category of autism and/or intellectual disability, (b) eligible for the state’s alternate assessment, (c) enrolled in at least one general education class, (d) receiving ongoing support from a paraprofessional who consented to participate in the study, and (e) on the caseload of a special education teacher who consented to participate in the study. Special educators were responsible for identifying students who met inclusion criteria whom they believed would benefit from peer support arrangements; students were not randomly selected. Based on student Individual Education Programs (IEPs), all four students scored in the first or second percentile on measures of intelligence and adaptive behavior.
Mark, Michelle, and Katie
Mark was an 11-year-old African American male with autism in the fifth grade. Mark’s primary form of communication was verbal speech. His IEP goals addressed basic literacy and writing (e.g., decoding words, writing a sentence, reading comprehension), math skills (e.g., adding and subtracting two- and three-digit numbers), language and communication, working independently, and asking for help.
Michelle, a special education paraprofessional, supported Mark in art class. Michelle was an African American female with an associate’s degree and 6 years of experience in her current role. Michelle had worked with Mark for 1 year. Michelle was supervised by Katie, a White female special education teacher with 25 years of experience in special education.
Brian, Angela, and Rachel
Brian was a 12-year-old White male with intellectual disability in the sixth grade. Brian’s primary form of communication was verbal speech in one to two word phrases. Brian’s IEP goals addressed basic math skills (e.g., matching manipulatives, identifying coins, telling time), pre-reading skills (e.g., ordering letters of his name, identifying sight words), expressive and receptive language, working on class activities, and making transitions with minimal prompting.
Angela, a special education paraprofessional, supported Brian during computer class. Angela was a White female with an associate’s degree and 12 years of experience in her current role. This was Angela’s first year working with Brian. Angela was supervised by Rachel, an African American special education teacher with a master’s degree and 8 years of experience in special education.
Joshua, Dave, and Ben
Joshua was a 13-year-old White male with intellectual disability and autism in the eighth grade. Joshua also had a mild hearing loss and a visual impairment. Joshua did not use verbal speech to communicate. Instead, he used a combination of gestures, body language, and vocalizations. Joshua’s speech therapist had introduced him to a picture communication book, but it was not utilized outside of structured speech therapy. Joshua’s IEP goals addressed functional math skills (e.g., identifying “same” and “different”), pre-literacy skills (e.g., identifying letter sounds), cleaning up materials, making transitions, following directions, and using a picture communication book to make choices or requests.
Dave, a special education paraprofessional, supported Joshua during art class. Dave was an African American male with a bachelor’s degree and 4 years of experience in his current position. He had worked with Joshua for 3 years. Dave was supervised by Ben, a White special education teacher with a master’s degree and 8 years of experience in special education.
Faith, Wanda, and Ben
Faith was a 10-year-old White female with autism who was in the fifth grade. Faith had a very limited verbal vocabulary that she used sparingly. Instead, she primarily used a combination of gestures, body language, and vocalizations to communicate. Her IEP goals addressed basic math skills (e.g., identifying numbers), pre-literacy skills (e.g., matching letters), self-care through toileting, and improving communication and fine motor skills.
Faith was supported during computer class by Wanda, an African American special education paraprofessional with a bachelor’s degree and 5 years of her experience in her current position. Wanda had worked with Faith for 5 years. She was supervised by Ben, the same special education teacher who supervised Dave.
Peers Without Severe Disabilities
Peers without severe disabilities participated in the study by agreeing to provide support to focus students. Special educators, general educators, and paraprofessionals worked together to identify peers they anticipated would work well with the focus student, would be willing to provide support, and might benefit themselves from providing support. Teachers were instructed to select peers who were enrolled in the same class as the focus student and did not have a severe disability. We encouraged teachers to consider peers with a history of positive interactions with the focus student, but this was not a requirement.
Twelve students were identified by teachers and paraprofessionals and given assent and consent forms. Two of these 12 students did not participate—one moved prior to the start of the intervention and the other did not have parental consent to participate. Three 10-year-old peers (one White male, one Hispanic male, and one Hispanic female) supported Mark during art class. One 13-year-old (Asian American female) and two 12-year-old peers (one White female and one African American female) supported Brian during computer class. Two peers (one 11-year-old Hispanic female and one 12-year-old White female) supported John during art class. Two peers, both 10-year-old Hispanic females, supported Faith during computer class.
School and Classroom Settings
All the participants attended a middle school in a large urban school district in the southeastern United States. The school served more than 700 students, and approximately half of the students were White. More than half of the students were eligible to receive free or reduced-price meals.
Computer class
Brian and Faith each attended a 75-min computer class. They were enrolled in different class periods taught by the same general education teacher. In each class students sat in rows of independent computer stations. There were 21 students in Brian’s class and 19 students in Faith’s class. According to observations across both students’ classes, 24.3% of observation intervals included large-group lecture, and 97.4% included independent work time. No small-group instruction was observed. (Each 10-min interval could include more than one instructional format.) In addition to supporting Brian, Angela also supported another student with a severe disability who was enrolled in the class. Faith was the only student with a severe disability enrolled in her class.
Art class
Mark and Joshua attended a 75-min art class where the intervention was implemented. They were enrolled in different class periods taught by the same general education teacher. Students sat at rectangular tables with 3 to 5 students at each table. There were 20 students in Mark’s class and 19 students in Joshua’s class. According to observations across both students’ classes, 78.2% of observation intervals included whole class lecture, 1.3% included small-group instruction, and 60.8% included independent work time. In addition to supporting Mark, Michelle also worked with another student with a severe disability who was enrolled in the class. Joshua was the only student with a severe disability enrolled in his class.
Experimental Design and Procedures
To evaluate the effectiveness of the intervention, we used a multiple-probe-across-participants design involving repeated intermittent measurement of the dependent variable with the introduction of the independent variable staggered across participants (Gast, Lloyd, & Ledford, 2014). Phase changes were made based on the primary dependent variable, an increase of interactions between focus students and peers.
Baseline condition
Before introducing the intervention, we conducted observations of typical classroom activities. During the baseline condition, all four participating students with disabilities received direct paraprofessional support. We instructed paraprofessionals to provide support to the focus students in the same manner as prior to the study. In art class, the general education teacher reported she had arranged for the focus students to work with peer tutors prior to the beginning of this project. During baseline observations, we observed the focus students sitting next to designated peer tutors, and the general education teacher sometimes instructed the peer tutors to work with focus students. Both the general education teacher and the paraprofessionals reported there was no formal process to design appropriate peer roles or orient peers to these roles. Both general education teacher and paraprofessional behavior related to encouraging focus students and peers to work together or interact during the baseline condition is described in Table 1.
Percentage of Observation Intervals With Paraprofessional Facilitation of Peer Support, Peer-Implemented Strategies, Interactions, and Proximity to Peers by Participant and Condition.
Note. Percentages reflect the average across all baseline or intervention observations sessions. — denotes that data were not collected for a variable in a given condition. Bl = baseline condition; Int = intervention condition; Gen = generalization condition.
Pre-intervention procedures
During the baseline condition, either the first or second author met with each supervising special educator for a 4-hr training focused on (a) the implementation steps associated with peer support arrangements and (b) how to deliver training that would enable paraprofessionals to implement these steps. During the training, we provided an intervention manual detailing the implementation steps (see Brock & Carter, 2015) for peer support arrangements and described and modeled each step. Similarly, we described and modeled each step of a training checklist. General educators were invited to attend a 20-min portion of this training so they could collaborate with the special educator and the researcher in identifying a target behavior for the focus student that was aligned with a goal from the student’s IEP and could be supported by peers during the class. In addition, the special education teacher and member of the research team identified a specific strategy to teach to peers to support the target behavior. Target behaviors and strategies for each focus student are defined in Table 2. Finally, the special educator worked with the general educator and paraprofessional to identify potential peers to support the focus student during the intervention condition.
Targeted Goals and Peer Strategies.
Note. Goals and strategies were developed in coordination with special and general educators, and were aligned with goals from the student’s Individualized Education Plan (IEP). To demonstrate correct use of a strategy, peers were required to follow all strategy steps until they reached the last step or the student demonstrated the target behavior.
Intervention condition
Special educators delivered a three-part training package to teach paraprofessionals to implement peer support arrangements. First, during an initial 2-hr training session, special educators explained the steps for implementing and facilitating the peer support arrangements, provided specific examples, and led the paraprofessional in creating a written peer support plan that outlined general support strategies as well as the individualized strategy developed in collaboration with the general educator and the researcher (see Table 2). Second, paraprofessionals viewed two researcher-created 10-min video models about facilitating social interactions and academic support among peers and the focus student. Third, 3 to 5 days after the initial training, special educators conducted a live 30-min observation in the classroom and then met with paraprofessionals to provide coaching and feedback. Teachers reinforced positive examples of paraprofessional facilitation and provided constructive feedback about how facilitation could be improved.
After completing their own training, paraprofessionals held an orientation with the two to three peers who were identified during the pre-intervention recruitment. Meetings took place during the focus class in an empty classroom. The first author attended all meetings to measure fidelity of implementation (see dependent variable section) and to explain and model the specific target behavior and strategy that was selected for the focus student. The paraprofessional began to provide ongoing facilitation and support for the peers in the general education class beginning on the first class meeting after the initial orientation. This facilitation included the strategies that were modeled and explained to them in their initial training and in the video models.
Generalization condition
Paraprofessionals were encouraged to continue implementing peer support arrangements when all four students transitioned from the targeted class (art or computer class) to a physical education class. The makeup of each class remained intact, so peers who were trained in the intervention condition remained in the same class as the focus student. No formal training or support was provided to the paraprofessionals by special educators or researchers during this condition in the new setting.
Teacher procedural fidelity
We measured teacher procedural fidelity with a 41-step checklist that captured key behaviors associated with delivery of training to the paraprofessional (available from the first author upon request). Katie independently implemented all 41 steps correctly with Michelle (100%); Rachel, 39 steps with Angela (95.1%); Ben, 40 steps with Wanda (97.6%); and 41 steps with Dave (100.0%). The first or second author attended all training activities and prompted teachers to implement any steps they failed to implement independently.
Dependent Measures and Recording
Orientation meeting with peers
The first author used a 10-step implementation checklist to measure the degree to which teacher-delivered training enabled paraprofessionals to implement with fidelity the steps of the initial training meeting with peers. Implementation steps included (a) providing a rationale for peer support arrangements, (b) describing background information about the focal student with severe disabilities, (c) describing general goals of peer support including promoting social interactions and academic engagement, (d) emphasizing the importance of confidentiality and respectful language, (e) setting expectations specific to the classroom (e.g., sitting with student, checking in with the paraprofessional), (f) outlining individualized peer support strategies from the written plan, (g) describing the individualized student goal and modeling the intervention strategy, (h) providing guidance on when to seek assistance from the general educator or paraprofessional, (i) soliciting and answering questions from the peers about their new roles, and (j) explaining that the peer support arrangement would begin with the next class meeting and that the seating arrangement would change to allow peers to sit next to the student with a disability. If a step was skipped or missed by the paraprofessional, the first author provided prompting or corrective feedback. Unprompted correct implementation of the orientation meeting was calculated as percentage of steps implemented correctly (before prompting or corrective feedback) out of the total number of steps.
Classroom observations
Measurement of dependent variables occurred during observations in the general education classroom two to five times each week across baseline and intervention conditions. Observers recorded data on paper forms and timed intervals using a smartphone application. Observations lasted for the entire class session.
Observers used a partial-interval recording system (10 s observe, 10 s record) to measure behaviors of the focus student and of peers. Focus student behaviors included interactions with peers and occurrences of the individualized target behavior. A focus student interaction was defined as the student directing verbal or nonverbal communicative behavior toward a peer without severe disabilities. Each focus student target behavior was identified and defined during pre-baseline procedures (see description above and Table 2). Peer behaviors included interacting with the focus student, prompting or reinforcing the focus student, and using the identified strategy. A peer interaction was defined as a peer without severe disabilities directing verbal or nonverbal communicative behavior toward the focus student. Peer reinforcement was coded when a peer praised the focus student, verbally or with gestures, for engaging in a desired academic or social behavior. Peers were coded as prompting the student with the disability when they encouraged or suggested a way for the focus student to engage in a desired academic or social behavior. Peers were coded as implementing the strategy correctly when they implemented all steps of the strategy selected for the focus student (see Table 2).
Observers also used the partial-interval recording system to measure paraprofessional behaviors to facilitate the peer support arrangements. Paraprofessional support was defined as the paraprofessional doing one or more of the following nine behaviors: (a) prompting social interactions between the focus student and peers, (b) reinforcing social interactions between the focus student and peers, (c) providing information aimed at helping peers interact socially with the focus student, (d) prompting academic support by encouraging peers and the focus student to work together in a specific way, (e) reinforcing academic support by praising the peers or focus student for the way they work together, (f) providing information to peers that might better help them academically support the focus student, (g) prompting proximity of the focus student and peers, (h) prompting peers to use the specific strategy for the focus student, and (i) checking in with peers to see if they are comfortable in their role or if they would like assistance. For examples of these behaviors, see Brock and Carter (2015). In addition, observers measured general educator behavior related to facilitating peer support, which was defined as any prompting or reinforcement directed toward the focus student or peers to encourage them to interact or work together. We measured general educator facilitation because one general educator reported already working to implement a peer-mediated intervention with two focus students, and we wanted to measure the extent to which she was promoting its implementation.
Observation Training and Interobserver Agreement
Observers were all graduate students studying special education. The first author led two 2-hr training sessions to review the training manual and provide verbal, written, and video examples and nonexamples of all codes. Before collecting experimental data, observers met the following criteria: 100% accuracy on a written test of coding definitions, at least 90% accuracy on all variables when coding three 10-min video recordings, and at least 90% agreement with an expert coder on all variables in a live setting. Brief follow-up trainings were provided when individualized student behaviors and peer strategies were defined, and observers were required to achieve at least 90% agreement with an expert coder in a live setting. A second observer collected data on 38.5% of all classroom observations, balanced across study participants and experimental conditions. Point-by-point agreement was calculated by dividing the number of intervals the primary and secondary observer codes matched by the total number of intervals. Across all primary dependent variables (i.e., all variables reported in Table 1 or graphed in Figure 1), average agreement ranged from 94.6% to 100%. Agreement never fell below 90.1% for any individual variable for any observation.
Percentage of intervals with paraprofessional behaviors facilitating peer support (triangles), interactions between students with disabilities and their peers (squares), peer interactions (open circles), and progress on individualized goals (closed circles).
Social Validity
After the end of the intervention condition, we gave peers a pencil-and-paper questionnaire asking them what they most enjoyed about providing peer support and if they would be willing to provide similar support in the future for students with disabilities.
Results
Functional relations were demonstrated between teacher-delivered training and paraprofessional implementation of peer support arrangements, paraprofessional implementation and peer support behavior, paraprofessional implementation and peer interactions, and peer support and student improvement on individualized goals. Teacher-delivered training enabled all four paraprofessionals to implement all components of peer support arrangements, all four focus students with disabilities experienced increased interactions with peers, all peers increased their delivery of support behaviors, and three of four focus students experienced improvement on individualized goals. Gaps in data collection for Mark (school days 26–31) and for Joshua (school days 25–32) were due to extended absences related to an illness. Outcome data are described below, as well as summarized in Figure 1 and Table 1.
Paraprofessional Implementation of Peer Support Arrangements
All four paraprofessionals successfully completed a peer support plan. Two paraprofessionals (Michelle and Dave) implemented all 10 steps (100%) of the initial meeting with peers correctly. Angela and Wanda each implemented nine steps correctly (90%). Both paraprofessionals did not clearly explain (without prompting) under what conditions peers should approach them for help. All four paraprofessionals increased their facilitation of peer support arrangements in the intervention condition, although data patterns and magnitude of change were varied (see Figure 1). Michelle and Angela did not demonstrate any facilitation behaviors during the baseline condition, and the level and variability of their facilitation increased immediately after receiving training, although the magnitude of this change seems small given the scaling of the graph required to display Dave’s facilitation behaviors, which were much more frequent. Dave demonstrated some facilitation behaviors during baseline, and there was a substantial increase in level immediately after training was introduced. Wanda only demonstrated one facilitation behavior during the baseline condition. Wanda’s facilitation behavior did increase in variability in the intervention condition, but there was not a clear and immediate change in level. Across all paraprofessionals, facilitation was more frequent for academic support compared with social interaction, and for prompting compared with reinforcement (see Table 1).
Peer Interactions, Peer Support Behaviors, and Progress on Student Goals
Peer interactions were either infrequent (for Mark and Joshua) or nearly absent (for Brian and Faith) during the baseline condition (see Figure 1). All four students experienced an immediate and substantial increase in the level of interactions after the introduction of peer support arrangements. For Brian and Faith, there was no overlap in level between the baseline and intervention conditions, but there was some overlap between conditions for Joshua and Mark. For Joshua, this overlap stemmed from an extreme data point early in the baseline condition. For Mark, intervention data had two downward trends, and the lowest points at the end of these trends overlapped with baseline data.
Similarly, peer prompting and reinforcement was either infrequent (for Mark and Joshua) or completely absent (for Brian and Faith) during the baseline condition (see Figure 1). Across all four students, the level of peer prompting and reinforcement increased immediately and substantially in the intervention condition. Peer use of individualized strategies was inconsistent across participants. Peers implemented the strategy across 85% of classes in the intervention condition for Mark, 66.7% for Brian, 100% for Joshua, and 100% for Faith. Average frequency of implementation within each class meeting is summarized in Table 1.
Three of the four students demonstrated progress on individualized goals after the introduction of peer support arrangements (see Figure 1). Mark’s performance on his individualized goal (sustained independent engagement) was variable in both the baseline and intervention condition, with no clear change between conditions. Brian did not demonstrate any sustained independent engagement during the baseline condition, and the variability and level of the data increased in the intervention condition. Joshua did not communicate using picture symbols at all during the baseline condition, and he did communicate using picture symbols during every class session in the intervention condition. Faith did not independently type on the keyboard in the baseline condition, and her typing increased immediately and substantially with the introduction of peer support arrangements.
Generalization of Paraprofessional Implementation, Peer Interactions, and Peer Support Behaviors
There was little evidence of generalization of paraprofessional implementation behaviors to the physical education setting (see Figure 1). Angela and Wanda demonstrated some facilitation behavior for at least one class session, but their behavior was inconsistent across class sessions. Michelle did not demonstrate any facilitation behaviors in the physical education setting. For Mark, Brian, and Faith, both peer interactions and peer prompting and reinforcement behaviors decreased substantially from the intervention condition but exceeded baseline levels (see Figure 1). In all three cases, the majority of interactions were with peers who had been trained as part of peer support arrangements in the previous setting (see Table 1). We were unable to collect generalization data for Joshua; Joshua’s parents requested that he not participate in physical education class due to a medical condition.
Teacher Facilitation of Peer Support
Mark and Joshua’s teacher reported she was already implementing peer tutoring before this experiment began. Our data show she was demonstrating behaviors related to facilitation of peer support for both Mark and Joshua in the baseline condition (see Table 1). For Mark (but not Joshua), this facilitation continued into the intervention condition.
Social Validity
Peers were asked whether they would be interested in providing peer support in the future and what the best thing was about being a peer support. When asked whether they would be interested in providing peer support in the future, 9 of 10 peers wrote, “yes.” When asked what the best thing was about being a peer support, peers indicated they enjoyed helping other people (seven responses), getting to learn about and getting to know the student with a disability (two responses), and carrying on a conversation with the student with a disability (one response).
Discussion
Peer support arrangements are a promising practice, but important questions remain about how to replicate and sustain this strategy with typical school resources. To inform our efforts to develop a sustainable and replicable model for implementation, we (a) tested the efficacy of promising teacher-delivered training to enable paraprofessionals to facilitate peer support arrangements and (b) measured both paraprofessional and peer implementation behaviors. As in previous studies, peer support arrangements resulted increased interactions between students with severe disabilities and their peers (cf. Carter et al., 2011; Shukla et al., 1999). Patterns of paraprofessional and peer behavior in the targeted and generalization setting contribute to a better understanding of implementation, and highlight opportunities for improved implementation through additional training and support.
Our findings reinforce prior research showing that teacher-delivered training enables paraprofessionals to implement peer support arrangements in ways that lead to improved social outcomes for students with severe disabilities (Brock & Carter, 2015). Under the direction of highly qualified teachers, paraprofessionals contributed to developing a peer support plan, invited peers, oriented peers to their roles, and provided ongoing facilitation to peers. Patterns of facilitation were similar to those in Brock and Carter, with relatively low but consistent rates of paraprofessional facilitation being sufficient to improve rates of interaction between the focus student and peers. Two patterns of paraprofessional facilitation are of particular interest. First, differences in rates of facilitation among paraprofessionals might reflect student and contextual factors. Paraprofessionals who supported students who primarily used nonverbal communication (Dave and Wanda) facilitated more frequently than those who supported students who were primarily verbal communicators (Michelle and Angela). Specifically, they more frequently prompted interactions, prompted academic support, provided information for social interaction, and provided information for academic support. Increased frequency of these types of facilitation might be necessary to facilitate arrangements between peers and students with more complex communication needs. Second, low rates of reinforcement across paraprofessionals suggest that paraprofessionals may need additional training related to positive reinforcement. Other studies have shown that many special educators deliver praise with surprising infrequency but that these rates can be increased substantially with targeted training (e.g., Reinke, Lewis-Palmer, & Martin, 2007; Sutherland & Wehby, 2001). The same targeted training strategies that have been effective with teachers—repeated performance feedback (Reinke et al., 2007) and self-evaluation (Sutherland & Wehby, 2001)—might also be effective with paraprofessionals.
In addition to replicating previous findings, this study extends the literature in three key ways. First, results provide insight into peer behavior within peer support arrangements, and the relationship between peer behavior and paraprofessional behavior. This is the first published study to report rates of peer prompting, reinforcement, and use of strategies targeting individualized goals. Data patterns show peers increased their rates of prompting and reinforcement after paraprofessionals oriented them to their new roles and provided facilitation. However, in some cases, rates of peer prompting and reinforcement decreased over time. Surprisingly, these decreases did not coincide with increased rates of paraprofessional facilitation, highlighting an opportunity to train paraprofessionals to more closely observe peer behavior and adjust the intensity of their facilitation as appropriate. In addition, the frequency with which peers implemented targeted strategies and the impact of these strategies on individualized goals varied across focus students. In general, peers had more difficulty implementing multistep strategies when the initial step required them to use a gestural prompt in the absence of a verbal prompt. Researchers have noted that special educators also often struggle with a tendency to overly depend on verbal prompts (Hume, Boyd, Hamm, & Kucharczyk, 2014). These findings suggest peers should either be provided extra training and support to implement multistep strategies that involve different types of prompts or asked to use simpler strategies with which they are more likely to be successful.
Second, results suggest both paraprofessionals and peers might need additional training, prompting, or support to implement peer support arrangements in a new setting. After transitioning to a new class, paraprofessional facilitation behavior, peer prompting and reinforcement, and focus student interactions decreased sharply. This lack of generalization highlights a need for additional paraprofessional training that is specific to the new setting. Perhaps a “booster” training session with the special education teacher, including the development of new peer support plan specific to the new setting, would have been more conducive to sustained implementation and improved student outcomes. Based on responses from most peers who indicated they enjoyed providing support and would be willing to continue providing support in the future, peers would likely have been responsive to paraprofessionals had they provided more facilitation. However, it is still notable that interactions were elevated above baseline levels and that the majority of these interactions were with peers who had been trained in the previous setting. Despite the reduction from levels in the trained setting, students may have experienced higher levels of interactions in the generalization setting than they would have if no intervention occurred.
Third, teacher-delivered training had a substantial impact even in settings where the general education teacher reported a peer-mediated intervention was already being implemented. In both art classes, the general education teacher reported peer tutoring was already being implemented, and baseline observations confirmed the general educator periodically did prompt peers to provide support and paraprofessional to supervise and provide assistance. Despite these efforts, paraprofessional facilitation, peer support behaviors, and focus student interactions were infrequent and inconsistent in the baseline condition. Only through focused teacher-delivered training were these behaviors increased substantially and maintained consistently. Based on these data, general educators who desire to implement peer-mediated interventions with paraprofessional support might be most successful if they team with special educators who can provide specialized training and support to the paraprofessional.
Implications for Practice
The results of this study have important implications for special educators. First, focused teacher-delivered training can enable paraprofessionals to implement promising practices such as peer support arrangements. For paraprofessionals working with students with severe disabilities, focused training is especially important to maximize the opportunities in a general education classroom for communicating, improving social skills, and building relationships. Before receiving training from special educators, paraprofessionals never or rarely encouraged students with disabilities and their peers to work with one another. Furthermore, based on measures of generalization in this study, follow-up training or coaching may be necessary to support paraprofessionals when they are expected to apply strategies to new settings or with new students. Second, special educators should use peer support arrangements to increase social interactions with peers and to promote progress on IEP goals. To effectively support this type of learning and growth, special educators should work with paraprofessionals to identify meaningful goals related to the general education class, plan for how peers can support the target behavior, and train peers to be successful in delivering their support. Third, school administrators should provide time and resources for special educators, general educators, and paraprofessionals to collaborate on interventions that improve the quality of inclusion for students with severe disabilities. In this study, implementation of peer support arrangements required time for special educators to train paraprofessionals; time for special educators, general educators, and paraprofessionals to work together to identify goals and plan supports; and time for special educators to observe paraprofessionals and deliver performance feedback.
Limitations and Future Directions for Research
Several limitations should be considered when interpreting results from this study, and these limitations provide directions for future research. First, while three of the focus students made notable progress on their individualized goals, one did not. Based on observational data, we hypothesize this lack of progress stemmed from peers having difficulty implementing a prompting hierarchy that began with only a gestural prompt before progressing to verbal prompts. In future studies, researchers might experimentally test whether prompting hierarchies are actually more difficult for peers to implement when they begin with a gestural prompt. Second, the frequency of paraprofessional facilitation was quite varied across paraprofessionals. We hypothesize these differences might be due to characteristics of the focus student (i.e., complexity of communication needs) that might necessitate more or less facilitation. In future studies, researchers might develop a priori hypotheses about which factors might necessitate different rates of facilitation, and then measure the degree to which these factors are actually predictive of facilitative behavior. Third, we highlighted specific aspects of paraprofessional implementation that might be improved through additional training, including reinforcement strategies and responsiveness to peer behavior. In future studies, researchers might test the type and intensity of training required to improve these aspects of implementation. Fourth, despite targeting independent academic engagement for Mark and Brian, levels in the intervention condition were lower than ideal. In future studies, researchers might explore how implementation might be modified to make larger increases in academic engagement in addition to social interactions. Fifth, students were selected by teachers and may not be representative of all students with severe disabilities. In future studies, researchers could choose students randomly from a pool of participants who meet criteria.
Conclusion
Promising interventions such as peer support arrangements have the potential to improve important outcomes for students with severe disabilities. However, these improved outcomes will only be realized with an improved understanding of implementation that enables practitioners to effectively replicate practices in schools. This study provides some initial insight into effective implementation and highlights several concerns about implementation issues that warrant further research. Identifying, studying, and solving issues of implementation could contribute to developing an improved and more powerful intervention.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Partial support for this research came from the Office of Special Education Programs, U. S. Department of Education, through Grant H325D100010 to Vanderbilt University.