JOURNALOFAPPLIEDBEHAVIORANALYSIS 2007, 40, 353–357 NUMBER2 (SUMMER2007) THE EFFECTS OF SCRIPTED PEER TUTORING AND PROGRAMMING COMMON STIMULI ON SOCIAL INTERACTIONS OF A STUDENT WITH AUTISM SPECTRUM DISORDER ANNA-LIND PETURSDOTTIR, JENNIFER MCCOMAS, AND KRISTEN MCMASTER UNIVERSITYOFMINNESOTA AND KATHY HORNER MINNEAPOLISPUBLICSCHOOLS This study examined the effects of scripted peer-tutoring reading activities, with and without programmedcommonplay-relatedstimuli,onsocialinteractionsbetweenakindergartnerwith autismspectrumdisorderandhistypicallydevelopingpeer-tutoringpartnersduringfreeplay.A withdrawal design with multiple baselines across peers showed no effects of peer tutoring on socialinteractions.Awithdrawaldesignwith1peerandcontinuingbaselinesacrosstheother2 peers showed that adding play-related common stimuli to the peer-tutoring activity increased socialinteractions during free play. DESCRIPTORS: autism, generalization, peer tutoring, programming common stimuli, socialinteraction _______________________________________________________________________________ One of the defining characteristics of chil- Anotherapproachthatofteninvolvesscripted dren with autism spectrum disorders (ASD) is interactions is peer tutoring, in which students impairedsocialinteraction.Poorsocialskillsare work together in pairs with reciprocal roles. related to negative outcomes such as rejection Peer tutoring has been found to increase bypeers,mentalhealthproblems,anddropping interactions of students with autism with out of high school (Pollard, 1998). Thus, typically developing peers during reading activ- increasing social interactions is an important ities in third grade (Kamps, Barbetta, Leonard, goal in the education of children with ASD. A & Delquadri, 1994). Peer incidental teaching number of researchers have used scripts suc- hasalsobeenfoundtoincreasepeerinteractions cessfully to promote social interactions (e.g., of preschoolers with autism (McGee, Almeida, Krantz&McClannahan,1993,1998;Sarokoff, Sulzer-Azaroff, & Feldman, 1992). However, Taylor, & Poulson, 2001). In these studies, Kamps et al. observed generalization of inter- interactions generalized across activities or actionsacrosssettings,whereasMcGeeetal.did settings after scripts were gradually faded. not. One approach to facilitating generalization is Thisresearchwasconductedinpartialfulfillmentofthe to ensure that ‘‘there are sufficient stimulus doctoral degree in educational psychology by the first components occurring in common in both the author at the University of Minnesota. We thank the participants in the study as well as their parents. We also training and generalization settings’’ (Stokes & thank Lesley Craig-Unkefer for her helpful suggestions Baer, 1977, p. 360). For example, program- and Molly McLoone for assistance in conducting the ming common stimuli increased generalization study. Correspondence can be addressed to Anna-Lind of question-asking skills of a 5-year-old child Pe´tursdo´ttir, Reykjavik City Department of Education, with autism (Esbenshade & Rosales-Ruiz, Frikirkjuvegi1,101Reykjavik,Iceland(e-mail:annalind. 2001). The purpose of the present study was [email protected]). doi:10.1901/jaba.2007.160-05 toexploretheeffectsofscriptedpeertutoringin 353 354 ANNA-LIND PETURSDOTTIR et al. reading activities, with and without pro- three target peers or they directed to him. grammed common play-related stimuli, on Interactions were considered directed to a peer social interactions between a young student if the speaker looked at or turned his head with ASD and his typically developing peer towards him at any time while emitting the tutors during free play that followed peer utterance, or if it was a response to something tutoring. said or done by a peer. Each separate play sound, laughter, and partial or whole sentence wascounted.Forexample,‘‘Don’ttake—;leave METHOD it!’’ counted as two interactions. Participants and Setting Two observers independently scored 33% of Max was a 5-year-old boy with ASD and free-play sessions from all phases from video- developmental delay. He attended a special tapes. Total agreement on dependent measures education class but also attended a general wascalculatedbydividingthesmallerfrequency kindergarten class. Max functioned within the count for each peer and category by the larger normalacademicrange.Despitefluentlanguage frequency count and multiplying by 100%. skills,herarelyinteractedsociallywithhispeers. Interobserver agreement averaged 98% (range, Three typically developing 5-year-olds from the 75% to 100%). kindergartenclass,Rick,Joe,andBob,whohad no history of playing with Max, were chosen Independent Variables as target peers because their teacher thought K-PALS activities included identifying letter they might work well with Max. Sessions took sounds, decoding, reading sight words, and place in the special education class, when readingsentences(Fuchs,Fuchs,Thompson,Al its 5 students were joined by about 10 kinder- Otaiba, Yen, McMaster, et al., 2001). During gartners. peer tutoring, Max and his peer took turns being the tutor, prompting the reader to Materials identify sounds and read words on a worksheet Participants used Peer-Assisted Learning using verbal prompts (e.g., ‘‘What sound Strategies for Kindergartners (K-PALS; Fuchs, [word]?’’ ‘‘Read it slowly.’’ ‘‘Sing it and read Fuchs, Thompson, Al Otaiba, Yen, McMaster, it.’’) and praise every 5 to 10 sounds. The tutor et al., 2001) worksheets during peer tutoring. systematically corrected errors (e.g., ‘‘Stop, that Programmedcommonstimuliweretoyswithat sound is—; what sound?’’). During a typical least 20 pieces that could be put together, such session, each student engaged in 30 to 60 as a marble run, that were typically available interactions, either as tutor or reader. Teaching during free play. Treatment integrity was staff provided prompts to stay on task, praised assessed with K-PALS checklists (Fuchs, Fuchs, on-taskandcooperativebehaviors,andawarded Thompson, Al Otaiba, Yen, Yang, et al., 2001) points that could be traded in for a reward. and a checklist to assess verbal prompts and When common play-related stimuli were pro- responses related to the common stimuli. grammed into the K-PALS activities, the tutor praised the reader after each line of text and Data Collection and Interobserver Agreement asked him to take one toy item and place it in Event recording was used to assess frequency an appropriate spot (e.g., one piece of a marble ofsocialinteractionsbetweenMaxandhispeers run on top of another). during 20-min free-play sessions immediately Treatment integrity. Average K-PALS integri- following reading sessions once to twice per ty was 81% (range, 56% to 92%) and 91% week. Social interaction was defined as any for common stimuli activities (range, 86% to verbal utterance Max directed to any of the 96%). Average interobserver agreement on PEER TUTORING AND SOCIAL INTERACTIONS 355 treatment integrity was 89% for K-PALS and parts of the same toy and were encouraged to 93% for common stimuli activities. Most continue playing with it. This encouragement integrity failures simply involved failure to was similar to the general prompts to play complete the entire K-PALS lesson. together (with specific toys or in certain play areas) provided to Max and the target peers Design and Procedure during free-play sessions in other phases. Max worked with each target peer on K- PALS activities during 15-min peer-tutoring Social Validity sessions, four times per week, for 4 to 5 Teaching staff completed a 14-item ques- consecutiveweeks.Duringthe20-minfree-play tionnaire on theacceptability, effectiveness, and periods, participants could choose among other factors related to programming play- a variety of activities (e.g., puzzles, art work, related stimuli into the reading activity. Each boardgames,andassemblytoys).Teachingstaff statement was evaluated on a 6-point Likert encouraged Max and the 3 target peers to play scale (1 5 strongly disagree; 6 5 strongly agree). together in the same area (e.g., at a table), but they were free to leave the designated area or RESULTS AND DISCUSSION playwithotherchildrenatalltimes.Noexplicit prompts or reinforcement for social interaction Figure 1 shows the frequency of social were programmed during any condition. interactions across baseline, scripted peer tutor- The effects of peer tutoring on social ing, and scripted peer tutoring with common interactions during free play were assessed with stimuli. The initial phases show that engaging a combined ABA withdrawal and multiple in scripted interactions during K-PALS reading baselineacross peers design. A CBC withdrawal activities did not lead to increases in social design was used to examine the effects of interactions between Max and any of the 3 programming common stimuli in the peer- typically developing peers during the subse- tutoring reading activity and free-play time on quent play sessions. These findings differ from the frequency of social interactions during free Kamps et al.’s (1994) observation that social play with 1 target peer (Rick) while baseline interactions increased during free time follow- measures continued forthe other 2target peers. ing academic peer-tutoring sessions, which Baseline (A). Baseline sessions consisted of might be due to several differences between 20-minfree-playperiodsimmediatelyfollowing the studies. Possibly, peer-tutoring tasks in reading sessions that occurred when Max and Kamps et al.’s study were less scripted than K- a target peer were not partnered for peer- PALS, providing more potential for social tutoring reading activities. interactions, making social interactions in other Scripted peer tutoring (B). These sessions settings more likely because only stimulus involved 20-min free-play periods immediately generalization was required, compared to both following reading sessions that occurred when stimulus and response generalization in this Max and a target peer were partnered for peer- study. Also, Kamps et al. included nonverbal tutoring reading activities (i.e., 4 to 5 consec- interactions when scoring social interactions, utive weeks per target peer). whereas we measured only verbal utterances in Scriptedpeertutoringwithcommonstimuli(C). this study. Moreover, Kamps et al. enhanced Thesesessionsinvolved20-minfree-playperiods positive interdependence among partners by immediatelyfollowingreadingsessionsinwhich assigning students to competing teams and play-related stimuli were programmed into the posting team points earned for cooperation, reading activities. After the reading activity was whereas the current study did not employ such completed, participants were given additional procedures. 356 ANNA-LIND PETURSDOTTIR et al. Figure 1. Frequency of social interactions between Max and his peers during 20-min free-play periods following baseline,peer tutoring, andpeertutoring withcommonstimuli. The observed lack of generalization of peer- in which they were taught and practiced, thus tutoring interactions to other settings is more not generalizing to settings that lacked common consistent with McGee et al.’s (1992) findings, stimuli. In the current study, verbal interactions in which incidental teaching by typically de- were likely under stimulus control of aspects of velopingpreschoolersincreasedtheirinteractions the reading activity (e.g., worksheets) that were with peers with autism during free play but not not present during free play, therefore necessi- during lunchtime. Possibly, verbal interactions tating added common stimuli across settings to were under tight stimulus control of the setting promote generalization. PEER TUTORING AND SOCIAL INTERACTIONS 357 When play-related common stimuli were REFERENCES programmed into Max’s and Rick’s peer- Esbenshade, P. H., & Rosales-Ruiz, J. (2001). Pro- tutoring activities, social interactions during gramming common stimuli to promote generalized free-play sessions increased markedly in each question-asking:Acasedemonstrationinachildwith autism. Journal of Positive Behavior Interventions, 3, phase of peer tutoring with common stimuli. 199–210. Even though Max was not partnered with Joe Fuchs, D., Fuchs, L., Thompson, A., Al Otaiba, S., Yen, and Bob during those phases, slightly increased L., McMaster, K., et al. (2001). Peer-assisted social interactions with them were observed. learning strategies: Kindergarten reading teacher manual. Retrieved May 4, 2005, from www. When play-related common stimuli were re- peerassistedlearningstrategies.net moved from the peer-tutoring activities, partic- Fuchs,D.,Fuchs,L.S.,Thompson,A.,AlOtaiba,S.,Yen, ipants’ social interactions during free play L., Yang,N.J.,etal.(2001).Isreadingimportantin decreased. Thus, programming play-related reading-readiness programs? A randomized field trial with teachers as program implementers. Journal of common stimuli into the peer-tutoring reading Educational Psychology, 93,251–267. activities led to modest increases in generaliza- Kamps, D. M., Barbetta, P. M., Leonard, B. R., & tion but not to maintenance of verbal interac- Delquadri, J. (1994). Classwide peer tutoring: An integration strategy to improve reading skills and tions from the academic setting to the free-play promote peer interactions among students with setting. Teaching staff found programming autismandgeneraleducationpeers.JournalofApplied common stimuli to be acceptable (M 5 6), Behavior Analysis,27,49–61. effective (M 5 6), and simple (M 5 6). Krantz, P. J., & McClannahan, L. E. (1993). Teaching children with autism to initiate to peers: Effects of The findings suggest that adding play-related ascript-fadingprocedure.JournalofAppliedBehavior commonstimulitoacademicpeertutoringmay Analysis,26,121–132. increase social interactions between young Krantz, P. J., & McClannahan, L. E. (1998). Social students with ASD and their peers. However, interaction skills for children with autism: A script- fading procedure for beginning readers. Journal of it should be noted that when play stimuli were Applied BehaviorAnalysis,31, 191–202. incorporated into the peer-tutoring activity, McGee, G. G., Almeida, M. C., Sulzer-Azaroff, B., & some of the reading time was spent playing, Feldman, R. S. (1992). Promoting reciprocal inter- actionsviapeerincidentalteaching.JournalofApplied resultinginslightlylesstimetopracticereading. Behavior Analysis,25,117–126. Also, time constraints (end of school year) Pollard, N. L. (1998). Development of social interaction allowed testing the effects of programming skills in preschool children with autism: A review of common stimuli with only 1 peer over a short the literature. Child & Family Behavior Therapy, 20(2),1–16. time period. Further research is needed to Sarokoff, R. A., Taylor, B. A., & Poulson, C. L. (2001). establish the generality of the effects with more Teaching children with autism to engage in conver- participants, more varied participants, and sational exchanges: Script fading with embedded across settings. Future research should also textual stimuli. Journal of Applied Behavior Analysis, 34,81–84. determine how long common stimuli need to Stokes, T. F., & Baer, D. M. (1977). An implicit be programmed in academic and play activities technology of generalization. Journal of Applied to achieve a high, stable rate of interactions Behavior Analysis,10,349–367. during free play and how common stimuli can Received November 1,2005 be faded without decreases in social interac- Final acceptanceDecember 22,2006 tions. Action Editor,Iser G. DeLeon