USING MULTIPLE-EXEMPLAR TRAINING TO TEACH A … · 2013. 8. 2. · Video Models The video models used to teach sharing depicted two peers sharing an activity shown from a third-person

USING MULTIPLE-EXEMPLAR TRAINING TO TEACH AGENERALIZED REPERTOIRE OF SHARING TO CHILDREN

WITH AUTISM

DENISE MARZULLO-KERTH, SHARON A. REEVE, AND KENNETH F. REEVE

CALDWELL COLLEGE

AND

DAWN B. TOWNSEND

INSTITUTE FOR EDUCATIONAL ACHIEVEMENT

The current study examined the utility of multiple-exemplar training to teach children withautism to share. Stimuli from 3 of 4 categories were trained using a treatment package of videomodeling, prompting, and reinforcement. Offers to share increased for all 3 children followingthe introduction of treatment, with evidence of skill maintenance. In addition, within-stimulus-category generalization of sharing was evident for all participants, although only 1 participantdemonstrated across-category generalization of sharing. Offers to share occurred in a novelsetting, with familiar and novel stimuli, and in the presence of novel adults and peers for allparticipants during posttreatment probes.

Key words: autism, generalization, multiple-exemplar training, social behavior, sharing, videomodeling

_______________________________________________________________________________

Prosocial behavior refers to a broad class ofresponses that includes smiling, cooperating,taking turns, making friends, expressing empa-thy, helping others, and sharing (Barton &Ascione, 1979; Bryant & Budd, 1984; Chan-dler, Lubeck, & Fowler, 1992; Cooke &Apolloni, 1976; Reeve, Reeve, Townsend, &Poulson, 2007; Rogers-Warren & Baer, 1976;Warren, Rogers-Warren, & Baer, 1976). Bryantand Budd (1984) suggested that sharing may bean especially important social behavior foryoung children because it increases opportuni-

ties for positive social interactions with peersand is an integral component of interactive play.

Several studies illustrate successful proceduresfor teaching children of typical developmenthow to share with their peers. Warren et al.(1976) for example, used modeling and rein-forcement to increase offers to share andcorresponding acceptances of share offers intwo groups of preschool children with reportedgeneralization to a new setting. Barton andAscione (1979) examined the generalization anddurability of sharing by teaching preschoolchildren to share verbally (e.g., ‘‘here yougo’’), physically (e.g., handing a toy), or both.Physical sharing increased for all childrenduring training, but it generalized and main-tained only for children who were taught toshare verbally. Verbal sharing increased duringtraining only for children who were taught toshare verbally or both verbally and physically;however, verbal sharing did not generalizeacross settings. Bryant and Budd (1984)extended this training package to six preschool-ers with disabilities, focusing on teaching

Correspondence concerning this article should beaddressed to Sharon A. Reeve, Department of AppliedBehavior Analysis, Caldwell College, 120 BloomfieldAvenue, Caldwell, New Jersey 07006 (e-mail: [email protected]).

doi: 10.1901/jaba.2011.44-279

This study was conducted as partial fulfillment of therequirements for the first author’s MA degree in appliedbehavior analysis from Caldwell College. Portions of thesedata were also presented at the 34th annual convention ofthe Association for Behavior Analysis, May, 2008, inChicago. We thank the students and staff at the Y.A.L.E.School for their participation.

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2011, 44, 279–294 NUMBER 2 (SUMMER 2011)

279

specific components of sharing responses: offers,requests, and acceptances. All three componentsof sharing increased after implementation ofinstructions, modeling, and behavior rehearsals.

Compared to children of typical develop-ment, children with autism often displayespecially marked impairments in sharing(Baron-Cohen, Leslie, & Frith, 1985; Eisenberg& Fabes, 1998; Rheingold & Hay, 1980;Rutter, 1978; Volkmar, Carter, Sparrow, &Cicchetti, 1993; Wing, 1988). Despite thesedeficits, few studies have examined methods toteach children with autism to share. Sawyer,Luiselli, Ricciardi, and Gower (2005) increasedverbal and physical sharing in one child withautism through the use of priming before playsessions, along with in-session prompts andreinforcement; however, they did not assess skillgeneralization. Recently, DeQuinzio, Town-send, and Poulson (2008) used a forwardchaining procedure to teach children withautism to approach a peer and emit a show-give-play sharing response chain with multipletoy exemplars. Sharing of toys increased for allparticipants and generalized to nontrained toys,novel locations, and different peers. However,DeQuinzio et al. conducted training and testingwith only one class of materials (toys), so it isnot clear whether sharing could be successfullytaught using other classes of materials orwhether sharing would generalize to untrainedclasses of materials.

Because the utility of a behavior-changeprocedure is increased when responding gener-alizes across persons, settings, and materials(Stokes & Baer, 1977; Stokes & Osnes, 1989),it is important to implement procedures thatpromote generalization. One strategy to accom-plish this involves the teaching of multipleexemplars. Multiple-exemplar training has beenused to teach numerous skills to individualswith moderate to severe disabilities, includingappropriate vending machine use (Sprague &Horner, 1984) and setting and clearing tables(Horner, Eberhard, & Sheehan, 1986). For

children with autism, multiple-exemplar train-ing has been used to teach helping behavior(Reeve et al., 2007), empathy (Schrandt,Townsend, & Poulson, 2009), and appropriateaffect (Gena, Krantz, McClannahan, & Poul-son, 1996), among others.

Stimuli that adequately reflect the diversity ofthe stimulus characteristics likely to be presentunder generalization conditions are used duringmultiple-exemplar training. The identificationof these stimuli by the teacher is known as ageneral case analysis (Engelmann & Carnine,1982; Horner, Sprague, & Wilcox, 1982;Sprague & Horner, 1984). When respondingcomes under the control of the stimulus featurespresent in the exemplars used during training,the presentation of novel stimuli that sharethese features should also occasion the sameresponse.

The purpose of the present study was toextend the findings of DeQuinzio et al. (2008)by establishing a generalized repertoire ofsharing in four children with autism. Amultiple-exemplar teaching procedure similarto that of Reeve et al. (2007) was used.Specifically, children were taught to share itemsfrom multiple classes of materials (art materials,snack foods, toys, and gym materials), andgeneralization was assessed both within classesof materials and across nontaught classes ofmaterials. In addition, teaching was conductedin different settings to promote generalizationacross locations, and discrimination of nonshar-ing situations was targeted and assessed. Finally,we assessed whether sharing was under appro-priate stimulus control by sampling situationsand stimuli for when sharing was appropriateand situations in which it was not. Also similarto Reeve et al., a combination of videomodeling and prompting was used for errorcorrection. Video modeling provided a meansto standardize the presentation of the modelsand has been shown to promote acquisition andgeneralization of skills for children with autism(Krantz, MacDuff, Wadstrom, & McClanna-

280 DENISE MARZULLO-KERTH et al.

han, 1991; LeBlanc et al., 2003; Nikopoulos &Keenan, 2004; Taylor, Levin, & Jasper, 1999).

METHOD

Participants

Four children with a diagnosis of autismparticipated. They attended a private school forindividuals with autism and had received theirdiagnoses through independent agencies priorto school enrollment. Direct observation andparent or teacher reports conducted prior to thestart of the study indicated that none of theparticipants engaged in sharing in the classroomor at home. Steven was 8 years 1 month old,Isaac was 7 years 10 months old, Bobby was8 years 1 month old, and Aiden was 7 years6 months old. Each participant had previousexperience learning skills with discrete-trialinstruction and using token-based motivationalsystems. According to reports by their teachers,each participant also demonstrated the prereq-uisite verbal skills needed to make the targetvocal responses used in the present study (e.g.,‘‘Want to try it?’’ ‘‘You try it’’), displayedcorrect imitation of various verbal and motorresponses using both in vivo and video models,and followed simple directions to engage inspecified activities.

Setting and Materials

Experimental sessions were conducted in aschool office that was not used for theparticipants’ daily education. The office con-tained a desk and three chairs, bookshelf, filingcabinet, computer, printer, and telephone.Additional materials included all relevant stim-ulus materials (i.e., toys, gym equipment, artmaterials, and snacks), a portable DVD player,small audio voice recorders, a token board, avideo camera with tripod, and data sheets.Approximately every 2 weeks, experimentalsessions were conducted in the school’s kitchento promote generalization. The kitchen con-tained two round tables with five chairs at each,stove, sink, refrigerator, dishwasher, and vend-

ing machine. Pre- and posttreatment sessionswere conducted in the usual daily classroom toassess generalization of sharing.

Materials and target responses for the studywere selected based on data collected with first-through third-grade general and special educa-tion students during a cumulative 2 hr ofobservations in public elementary schoolsconducted during free play, snack time, outdoorrecess, and gym. Materials used includedmodeling clay, crayons, stamps, jump rope,balls, puppets, pretend food, blocks, smallplastic animals, books, cars, and snacks. Themajority of the observed sharing statementswere requests or directions to share anotherstudent’s materials (e.g., ‘‘want one?,’’ ‘‘hereyou go,’’ ‘‘wanna play?,’’ ‘‘your turn,’’ and‘‘here’’). The classroom teachers were also askedwhat types of materials the participants werelikely to use during free time and group time,when offering to share would be appropriate.

The four stimulus categories were art mate-rials, snack foods, toys, and gym materials.Stimulus categories and items assigned to eachparticipant are listed in Table 1. In each ofthese four categories, five different materialsserved as multiple exemplars of the samestimulus category. For example, art materialsincluded crayons, dot paint, markers, coloredpencils, and glitter pens. Snack foods for Steven,Bobby, and Aiden included pretzels, cookies,chips, candy, and marshmallows. Specific snackfoods were different for Isaac because hedisplayed disruptive behavior in response tothe removal of the original snacks (which werehighly preferred) during the initial treatmentsession. To reduce disruption, snack foods forIsaac included carrots, celery, apples, andraisins. During pre- and postintervention gen-eralization probes, however, snack stimuli forIsaac were the same as those used for the otherthree participants. Toys included Lego blocks,cars, magnets, Play-Doh, and Peg-Board. Gymmaterials included a Velcro mitt and ball,basketball, scooter, Hippity-Hop, and Velcro

TEACHING SHARING 281

darts. One exemplar in each category wasrandomly selected for use in pre- and posttreat-ment generalization probes only (i.e., glitterpens, marshmallows, peg board, and Velcrodarts). In addition, materials that served asdiscriminative stimuli for nonsharing responseswere present, including papers to put in abackpack, an academic worksheet, books on thefloor that needed to be placed on a table, a dirtytabletop and a towel, and an article of clothing.It should be noted that although the materialsused for sharing and nonsharing trials were allselected based on observations of peers in

natural settings, no assessment was maderegarding the participants’ relative preferencesfor any of the materials.

Video Models

The video models used to teach sharingdepicted two peers sharing an activity shownfrom a third-person viewpoint (i.e., sceneperspective). Specifically, a 7-year-old boy wasfirst shown either sitting at a table or standing,engaged in an activity. A second 7-year-old boythen came on camera, approached the first boy,and stood within 0.6 m of him. Following the

Table 1

Assignment of Stimuli and Verbal Responses Across Participants

Steven Isaac Bobby Aiden

Teaching Art:crayonsdot paintmarkers

Snack:celeryapplesraisins

Toys:magnetsPlay-DohLego blocks

Gym:Hippity-HopVelcro mitt and ballmagnetic darts

Snack:pretzelscookieschips

Toys:carsmagnetsPlay-Doh

Gym:scooterHippity-HopVelcro mitt and all

Art:crayonscolored pencilsdot paint

Toys:Lego blockscarsmagnets

Gym:basketballscooterVelcro mitt and ball

Art:markerscolored pencilscrayons

Snack:candypretzelscookies

Within-categoryprobe

Art: colored pencilsSnack: candyToys: Play-Doh

Snack: carrotsToys: Lego blocksGym: Hippity-Hop

Art: dot paintGym: basketballToys: cars

Art: markersSnack: chipsGym: scooter

Across-categoryprobe

Gym: scooter Art: crayons Snack: chips Toys: cars

Verbal responses ‘‘Would you like to try this?’’‘‘Do you want to try?’’‘‘Here, you try it.’’‘‘Try this.’’

‘‘Try this.’’‘‘Do you want to try?’’‘‘Here, you try it.’’

‘‘Why don’t you try?’’‘‘Try this.’’‘‘Do you want to try?’’

‘‘Here, you try it.’’‘‘Would you like to try this?’’‘‘Do you want to try?’’

Snack:marshmallows (G)candy (P)cookies (T)

Snack:marshmallows (G)candy (P)pretzels (T)

Snack:chips (P)marshmallows (G)pretzels (G)

Snack:marshmallows (G)chips (P)candy (T)

Pre- andposttreatmentgeneralizationtrialsa

Art:glitter pens (G)colored pencils (P)markers (T)

Art:crayons (P)glitter pens (G)markers (G)

Art:glitter pens (G)dot paint (P)crayons (T)

Art:glitter pens (G)markers (P)dot paint (T)

Gym:scooter (P)basketball (G)Velcro darts (G)

Gym:Velcro darts (G)Hippity-Hop (P)Velcro mitt and ball (T)

Gym:Velcro darts (G)basketball (P)Hippity-Hop (T)

Gym:Velcro darts (G)magnetic darts (P)Hippity-Hop (T)

Toys:Peg-Board (G)Play-Doh (P)Lego blocks (T)

Toys:Peg-Board (G)Lego blocks (P)cars (T)

Toys:Peg-Board (G)cars (P)Play-Doh (T)

Toys:cars (P)Peg-Board (G)magnets (G)

a Includes stimuli used only during pre- and posttreatment generalization trials (G), probe trials during experimental

sessions (P), and teaching sessions (T).


approach of the second boy, the first boy heldout his item and emitted a verbal offer to share.Six video models were used for each participant,and each video model was approximately 8 to10 s in duration. Stimuli featured in each videomodel corresponded to stimuli used duringtraining trials for that particular participant.Similarly, verbal offers to share emitted by thepeer in the video model were verbal responsesassigned to that particular participant duringtreatment. Table 2 provides a list of the stimuliand responses featured in the video models foreach participant. For example, one of Steven’svideo models depicted a peer coloring withcrayons who asked, ‘‘Do you want to try?’’ inresponse to the approach of the second peer.

Response Measurement

Trained observers scored participant respons-es on a trial-by-trial basis using pencils andpaper data sheets. A correct offer to share wasscored only if both the motor and verbalcomponent occurred. A correct motor responseconsisted of the participant holding out an itemto the experimenter within 5 s of her approachto within 0.6 m of the participant. A correctverbal response consisted of a request for theexperimenter to engage in an activity with thepresented item (e.g., ‘‘Do you want to try?’’;‘‘Here, you try it’’) or an approximation of theresponse (e.g., ‘‘Do you want it?’’; ‘‘try’’). A

correct nonsharing response was scored if thechild responded on a nonsharing trial bycompleting the task without offering to share,either physically or verbally. Table 2 depicts therandomized assignment of verbal responsestaught to participants.

A teacher with master’s level training inbehavior analysis served as a second observerand scored sharing and nonsharing either invivo or via videotapes of a session. For eachtrial, the scoring of the two observers wascompared, and only an exact match wasconsidered an agreement. Percentage agreementwas calculated by dividing the number ofagreements by the number of agreements plusdisagreements and multiplying by 100%.Interobserver agreement data were collected onthe dependent variable during 54% of Steven’ssessions, 44% of Isaac’s sessions, 47% ofBobby’s sessions, and 40% of Aiden’s sessions.The mean percentage of agreement across alltrial types (training, probe, and nonsharingtrials) and across all participants was 99%(session range, 97% to 100%). A trainedobserver also collected data on the accuracy ofthe experimenter’s presentation of discrimina-tive stimuli, reinforcement, trial types, andvideo model error correction. These proceduralintegrity data were collected during 95% of allbaseline, treatment, maintenance, and pre- andposttreatment sessions. The mean percentage

Table 2

Video Model Types Across Participants

Video model

Verbal response Stimulus sets

Steven ‘‘Would you like to try this?’’‘‘Do you want to try?’’‘‘Try this.’’

cookies, carsmarkers, crayonsmagnets, chips

Isaac ‘‘Try this.’’‘‘Do you want to try?’’‘‘Here, you try it.’’

magnets, chipsVelcro mitt and ball, pretzelball, Play-Doh

Bobby ‘‘Why don’t you try?’’‘‘Try this.’’‘‘Do you want to try?’’

colored pencils, Velcro mitt and ballPlay-Doh, scootermagnets, markers

Aiden ‘‘Here, you try it.’’‘‘Would you like to try this?’’‘‘Do you want to try?’’

football, colored pencilscookies, carspretzels, Velcro mitt and ball


accuracy on procedural components was 99%(session range, 98% to 100%). In addition, asecond observer scored procedural integrity for46% of Steven’s sessions, 34% of Isaac’ssessions, 36% of Bobby’s sessions, and 34% ofAiden’s sessions. The mean percentage ofagreement on the accurate presentation ofprocedural components was 99% (sessionrange, 98% to 100%).

Design and Assignment of CategoriesA concurrent multiple-probe design across

participants was used. Three of the four possiblestimulus categories were assigned to the trainingcondition for each participant; the fourthcategory was used to assess across-categorygeneralization of sharing responses (see Ta-ble 1). Each category contained five possiblestimulus exemplars: Three were directly target-ed as stimulus exemplars (partially counterbal-anced across participants), a fourth exemplarwas used to assess within-category generaliza-tion of sharing during training sessions, and thefifth exemplar was used to further assess within-and across-category generalization during thepre- and postintervention generalization probes.For example, Steven’s teaching trials includedcrayons, dot paint, and markers (art materialscategory); pretzels, cookies, and chips (snackfoods category); and Lego blocks, cars, andmagnets (toy category). Probe trials for Stevenincluded a scooter (gym materials category) toassess across-category generalization, coloredpencils to assess within-category generalizationfor art materials, candy to assess within-categorygeneralization for snack foods, and Play-Doh toassess within-category generalization for toys.

ProcedureEach baseline and treatment session consisted

of 18 trials. Of these, nine were teaching trials(three exemplars from each of the three trainingcategories). Four additional trials were general-ization probes with the fourth exemplar fromeach of the three training categories and oneexemplar from the fourth stimulus category. To

promote response generalization, three of thefive experimenter-defined verbal offers to sharewere randomly assigned to each participant andwere rotated across trials. At the start of eachtrial, the experimenter approached and stoodwithin 0.6 m of the participant while he wasengaged in a specific activity. The purpose ofpresenting these discriminative stimuli was toset the occasion for the participant to offer toshare the item or items with which he wasengaged. Finally, five nonsharing discrimina-tion trials were interspersed among the nineteaching and four probe trials. To reducepotential order or sequence effects, the orderof the trials was randomized, with the exceptionthat all sessions began and ended with a trainingtrial.

Baseline. The participant was seated at a desk.The experimenter presented stimulus materialsfor that trial (e.g., crayons and a blank coloringpage) with a verbal direction to engage in thetarget activity (e.g., ‘‘color the picture’’). Within5 s of presenting the materials, the experimenterapproached the participant and stood within0.6 m of him. If the participant emitted acorrect verbal and motor offer to share, theexperimenter responded ‘‘sure’’ or ‘‘thanks,’’took the offered item, and manipulated theitem appropriately. No prompts or models wereprovided, and no tokens were delivered forsharing. The experimenter delivered tokens on avariable-ratio (VR) 2 schedule for following thedirection to engage in the task and attending tothe experimenter. The experimenter did notrespond if the participant emitted only a correctmotor response or a correct verbal response. Ifthe participant did not emit a correct responsewithin 5 s of the presentation of the discrim-inative stimuli, the experimenter removed thematerials and walked away. After a 5-s intertrialinterval, the experimenter presented the partic-ipant with the next activity and a new trialbegan. At the conclusion of a session, partici-pants could trade in their tokens for previouslyselected activities or materials (e.g., access to a


video game, preferred snacks or toys) that werenot used during the sessions.

Treatment. The steps of the procedure aredepicted in Figure 1. As in baseline, theexperimenter presented the participant withstimulus materials, gave the direction to engagein the target activity, waited 5 s, approached theparticipant, and stood within 0.6 m of him.Following a correct sharing response duringtraining trials, the experimenter accepted theoffered item; confirmed the offer to share bysaying ‘‘yes,’’ ‘‘sure,’’ or ‘‘thank you’’; manip-ulated the offered item appropriately; and thendelivered a token. After appropriately manipu-lating the item and returning it to theparticipant, the experimenter walked away andgave the direction to clean up. Materials wereremoved from the table and the trial ended. Anew trial began following a 5-s intertrialinterval.

If the participant did not respond within 5 sof presentation of the discriminative stimuli,responded incorrectly, or emitted only onecomponent of the sharing response (e.g., motorresponse or vocal response alone), the experi-menter removed the materials and implementedan error-correction procedure similar to thatdescribed by Reeve et al. (2007). First, theexperimenter presented a video model of ascenario in which appropriate sharing wasdepicted. Video models were used only whena participant failed to emit appropriate sharingbehavior during training trials. For ease ofimplementation and to promote generalizationof sharing, the video models did not necessarilyshow the same activity in which the participantwas engaged during that trial. For example, avideo model depicting a peer eating cookies andasking ‘‘Would you like to try this?’’ could bepresented to Steven during a trial in which hewas playing with cars and the target responsewas ‘‘Would you like to try this?’’

None of the participants required promptingto watch the video. When the video modelended, the experimenter presented the materials

for that trial a second time and approached andstood within 0.6 m of the participant. If theparticipant offered to share within 5 s, theexperimenter accepted the offered item; re-sponded with ‘‘yes,’’ ‘‘sure,’’ or ‘‘thank you’’;manipulated the item appropriately; and deliv-ered a token to the participant’s token board. Ifthe participant did not offer to share within 5 sof the second opportunity to do so, theexperimenter provided a physical prompt byusing hand-over-hand guidance to assist thechild in handing the items with which he wasengaged to the experimenter. Simultaneously,the experimenter activated a voice recorder thatemitted a recording of the target vocal response.If the participant emitted only one componentof the response (e.g., physical or vocal), theexperimenter delivered the correspondingprompt for the absent component. For example,if the child handed an item to the experimenterbut did not emit a target vocal response, theexperimenter provided only an audio promptby activating the voice recorder. If the childemitted a target vocal response but did notphysically offer to share, the experimenterprovided only a physical prompt with hand-over-hand guidance. The physical and auditoryprompts were repeated until the participantemitted the correct response independently.

As noted previously, generalization probetrials and nonsharing trials were interspersedwith training trials. During probe trials, theprocedure was identical to that during baseline.During nonsharing trials, the experimenterpresented the nonsharing stimuli to the partic-ipants and emitted an instruction appropriatefor those stimuli (e.g., ‘‘Can you please put thebooks on the shelf?’’; ‘‘Can you please wipe thetable?’’; ‘‘Put the papers in your backpack’’).When the participant completed the nonsharingtask (e.g., placed papers in a backpack), theexperimenter delivered a token and the trialended. If a child emitted a sharing response(e.g., ‘‘Want to try?’’) during a nonsharing trial,the experimenter paused for 5 s and then


Figure 1. Instructional procedure for sharing trials during treatment.


presented the same nonsharing trial a secondtime while verbally instructing the child tocomplete the task. This continued until thechild completed the task and did not emit asharing response. These nonsharing trials wereincluded to ensure that the participants attend-ed to the relevant discriminative stimuli that setthe occasion to offer to share and to preventovergeneralization of the target response tononsharing stimuli.

One participant (Isaac) engaged in disruptivebehavior during the general prompting proce-dure during training trials. To facilitate skillacquisition and reduce the occurrence ofdisruption, additional prompts were usedduring treatment. A gestural prompt was addedto the discriminative stimuli in each trial, and abehavior-specific praise statement was deliveredfor offers to share. During teaching trials, afterpresenting the stimulus materials and givingIsaac a verbal direction to engage in the activity,the experimenter approached Isaac, stoodwithin 0.6 m, and extended an open palmtoward him. Following errors, the video modeland auditory or manual prompts were deliveredas in the treatment procedure. In addition,when Isaac shared correctly, the experimenteremitted a behavior-specific praise statement(e.g., ‘‘Great! You handed it to me and said,‘try this.’’’). The experimenter no longerdelivered behavior-specific praise when Isaacmet the mastery criterion and no longerdelivered the gestural prompt after the scheduleof reinforcement was thinned.

The mastery criterion was offering to shareindependently on at least 89% (eight of nine) ofthe teaching trials for four consecutive sessions.The fixed-ratio (FR) 1 schedule of tokendelivery for correct responding was thinned toVR 2 when the participant met the masterycriterion. After criterion-level responding oc-curred during teaching trials for two consecutivesessions at VR 2, the schedule was thinned toVR 5. After two consecutive sessions ofcriterion-level responding on the VR 5 sched-

ule, the posttreatment generalization probetrials were conducted. The schedule of rein-forcement for Aiden was not systematicallythinned due to his completion of participationat the end of the school year.

Pre- and posttreatment generalization probes.The experimenter conducted probes assessinggeneralization of sharing to novel settings, novelpeople, and novel stimuli during two sessionsprior to baseline and two sessions after thetraining criterion was met. The 12 trials in eachsession consisted of three trials of teachingstimuli, four trials of probe stimuli that servedas within- and across-category generalizationprobes during training sessions, three trials ofthe fifth stimuli from each of the three trainingcategories that were not presented duringtraining, and two trials of additional stimulifrom the fourth, nontraining category. With theexception of the scooter, Hippity-Hop, dotpaint, and magnets, stimuli used in the pre- andposttreatment generalization probes were neverpresented in experimental sessions. Table 1depicts stimulus types used in pre- andposttreatment generalization probe trials foreach participant. During these probe sessions, anovel instructor presented the stimulus materi-als and gave the direction to engage in the targetactivity. In addition, 5 s after the presentationof the materials, a peer was directed to ‘‘seewhat [participant’s name] is doing’’ and toapproach the participant and stand within 0.6 mof him. The peer, who was approximately thesame age as the participant, was another childwith autism. This trial format was uniquebecause it was a peer who approached theparticipant rather than the person who initiallypresented the materials. The participant did notreceive tokens for correct responses or on-taskbehavior, nor did he receive prompts or videomodels following incorrect responses. Noadditional prompts were provided if theparticipant emitted a correct offer to share,and the peer accepted the item. If theparticipant emitted a correct response and the


peer did not take the item, the instructordirected the peer to accept the item and tomanipulate it appropriately (e.g., ‘‘You can eatthe chip,’’ or ‘‘Take the crayon and color withSteven.’’). The materials were removed and thetrial ended if the participant did not emit acorrect response within 5 s of the presentationof the discriminative stimuli.

Maintenance. Maintenance sessions wereconducted 1, 2, and 5 weeks after theposttreatment generalization probes for Steven,and 1, 2, and 3 weeks after the posttreatmentgeneralization probes for Isaac and Bobby.Maintenance sessions were not conducted forAiden because the school year ended and he wasno longer available to participate. Procedureswere identical to those in baseline except thatthe experimenter did not deliver tokens for on-task behavior.

RESULTS

Figure 2 shows the percentage of teaching,probe, and nonsharing trials in which eachparticipant independently offered to share,across consecutive sessions. Probe data arecollapsed across both within-category andacross-category generalization trials. Duringbaseline, none of the participants offered toshare during any trial, with the exception ofAiden, who handed a nonpreferred food item tothe experimenter on two occasions and said‘‘here.’’ Following the successive introduction oftreatment across participants, systematic in-creases in offering to share occurred duringboth teaching and probe trials. The percentageof trials in which Steven offered to share duringteaching trials systematically increased from 0%during baseline to 100% during treatment.During probe trials, the percentage of trials inwhich he offered to share increased from 0%during baseline to 75% during treatment.Isaac’s offers to share during teaching trialschanged in a similar manner, increasing from0% during baseline to 100% during treatment.His offers to share during probe trials increased

from 0% during baseline to 75% duringtreatment. Bobby demonstrated an increase inoffers to share on teaching trials from 0%during baseline to 100% during treatment andfrom 0% during baseline to 50% duringtreatment for probe trials. Aiden also demon-strated an increase in offers to share on teachingtrials from a mean of 2% during baseline to89% during treatment, and from 0% duringbaseline to 75% during treatment for probetrials.

Higher percentages of correct responsesoccurred during within-category probe trials(Steven, M 5 80%; Isaac, M 5 67%; Bobby,M 5 63%; Aiden, M 5 60%) than duringacross-category probe trials (data not shown).Specifically, Steven offered to share duringacross-category generalization probe trials dur-ing only one treatment session (M 5 9%).Bobby never demonstrated across-category gen-eralization during treatment, and Aiden dem-onstrated generalization during two treatmentsessions (M 5 13%). Isaac was the onlyparticipant who demonstrated across-categorygeneralization on several occasions (M 5 56%).

Throughout experimental sessions, all partic-ipants displayed some degree of vocal responsegeneralization beyond the three scripted targetphrases they were taught. Steven made nineunscripted vocal offers to share, Isaac madeseven, Bobby made two, and Aiden made seven.Examples of nontaught offers to share included‘‘Would you like to draw?’’; ‘‘Would you like tobuild?’’; ‘‘ride the scooter’’; ‘‘try, please.’’

None of the participants offered to shareduring any trial presented during the twopretreatment generalization probe sessions (Fig-ure 2). After each participant met the masterycriterion, however, the percentage of trials inwhich each participant offered to share in thepresence of both familiar and novel stimuli, in anovel setting, and with novel adults and peersincreased compared to the pretreatment gener-alization probes. Steven offered to share with apeer on 100% of the trials presented for both


Figure 2. Percentage of trials with correct independent offers to share across consecutive sessions.


posttreatment sessions. During the first post-treatment probe, Isaac’s offers to share with apeer increased to 58% of trials and to 92% oftrials during the second probe. Bobby did notoffer to share during his first posttreatmentgeneralization sessions, and offers to share weremoderately high (58%) during the secondposttreatment session. He did, however, phys-ically offer to share (holding out an item to hispeer) on 75% of trials during the firstposttreatment session and 92% of trials duringthe second session. His data demonstrate anincrease in physical sharing with his peers fromphysically offering to share during 0% of allpretreatment trials. Posttreatment generaliza-tion sessions were not conducted with Aidendue to time limits. Across the trials thatassessed generalization, Isaac and Bobby hadhigher percentages of correct responses onwithin- and across-category probe trials thatfeatured familiar stimuli that were not associ-ated with teaching than they did on within-and across-category trials that featured novelstimuli. Steven maintained criterion levels ofresponding during maintenance probes whenno tokens were delivered. Isaac’s offers to shareduring maintenance were just below thecriterion at 78% (seven of nine) of trials.Bobby offered to share during 100% ofmaintenance trials.

During baseline, none of the participantsoffered to share during nonsharing trials (i.e.,trials with task stimuli rather than sharingstimuli). As the treatment was introduced, offersto share during nonsharing trials increasedsomewhat for all participants: up to 20% forSteven, Bobby, and Aiden, and up to 40% forIsaac. Steven’s, Bobby’s, and Aiden’s offers toshare during nonsharing trials then decreased to0% and remained at 0% during the last fivetreatment sessions. During maintenance probesessions, Steven did not offer to share nonshar-ing stimuli, and Isaac and Bobby did not offerto share nonsharing stimuli during two of thethree maintenance probes.

DISCUSSION

Four children with autism learned to offer toshare in the presence of appropriate discrimi-native stimuli during teaching trials. In addi-tion, all participants demonstrated stimulusgeneralization by offering to share during probetrials using materials from the same category ofstimuli used during teaching and, for oneparticipant, from categories that were not usedduring teaching. Pre- and posttreatment gener-alization probes indicated that offers to shareincreased for three participants in a novelsetting, with a novel instructor and peer, andin the presence of novel and familiar stimulifollowing treatment. Sharing occurred infre-quently when it was not contextually appropri-ate (i.e., nonsharing trials). In addition, Steven,Isaac, and Bobby maintained the skill ofoffering to share during maintenance sessionsfollowing the termination of treatment. Finally,all participants demonstrated some responsegeneralization by emitting unscripted vocaloffers to share. Thus, the current studydemonstrated the establishment of a generalizedrepertoire of offering to share in these fourchildren with multiple categories of stimulusmaterials.

The present study extends the findings ofDeQuinzio et al. (2008) in a number of ways.First, sharing was taught across multiplestimulus categories, which likely producedrobust within-category generalization of sharing(but a lesser degree of across-category general-ization). In addition, teaching multiple exem-plars of vocal responses appropriate for sharingalso likely produced the vocal response gener-alization of sharing. By including trials in whichit was not appropriate to share, we demonstrat-ed that sharing had come under the control ofrelevant stimulus characteristics and that shar-ing infrequently or rarely overgeneralized tononsharing scenarios. Finally, maintenance ofsharing was observed after reinforcement wasthinned at specified criterion levels of respond-ing.


During teaching trials in which no responseor an incorrect response occurred, the presen-tation of the video model correction procedurepreceded a correct response during the subse-quent presentation of the in vivo discriminativestimuli during only 31% of trials acrossparticipants (data not shown). The second partof the error-correction procedure (additionalauditory or manual prompts) was used with thevideo model for the other 69% of training trialsin which no response or an incorrect responseoccurred. These data suggest that the combina-tion of a video model with audio and manualprompts may have been more effective inproducing the target response than was thevideo model alone. A future study mightcompare the effects of these error-correctionprocedures under controlled conditions.

Prosocial behavior will be functional forchildren with developmental disabilities onlywhen the behavior generalizes beyond trainingconditions across responses, stimuli, locations,and individuals (Reeve et al., 2007; Stokes &Baer, 1977). Behavior analysts can use a generalcase analysis when designing teaching strategiesfor generalized prosocial behavior by firstidentifying the relevant discriminative stimulithat occasion appropriate responses undernatural conditions, and then ensuring that thesecharacteristics are present under teaching con-ditions (Engelmann & Carnine, 1982; Horneret al., 1982; Sprague & Horner, 1984). In thepresent study, this was accomplished byobserving children sharing in their naturalsettings. Once identified, these putative dis-criminative stimuli for sharing were groupedinto categories for training purposes. Multipleexemplars were presented from each of thestimulus categories assigned to teaching trials toprogram for generalization across novel stimuliwithin the same stimulus category (Reeve et al.,2007). It is likely that generalization of sharingoccurred across novel stimuli from the samecategory because the set of training exemplarsadequately reflected the diversity of the stimulus

characteristics that were present under general-ization conditions. As a result, the outcomes ofthe present study add to the previous researchliterature that has shown that multiple-exemplartraining is effective for teaching numerousgeneralized skills (Gena et al., 1996; Horneret al., 1986; Reeve et al., 2007; Schrandt et al.,2009; Sprague & Horner, 1984).

Response generalization of verbal offers toshare was also assessed in the present study.Three verbal responses were assigned to eachparticipant for training. During treatment,Steven, Isaac, Bobby, and Aiden each learnedto emit all three of their target statements.However, the participants emitted a variablenumber of novel verbal offers to share. Futurestudies may determine whether additionalresponse exemplar training might produce agreater degree of response generalization andmore consistency across participants.

Generalized sharing across categories was notobserved for Steven and Bobby, and limitedgeneralization was displayed by Aiden. Theseoutcomes may be attributed to the singlepresentation of an across-category generaliza-tion probe during each treatment session. It ispossible that additional opportunities for re-sponding to across-category probe stimuliwould have produced somewhat higher per-centages of trials in which sharing occurred.Given that Isaac did demonstrate across-cate-gory generalization of sharing, it is also possiblethat the features of the across-category probestimuli for Isaac (art category, crayons) weremore similar to the features of the trainingstimuli from his training categories than wereSteven’s (gym category, scooter) and Bobby’s(snack category, chips). For example, a group ofmany crayons was presented to Isaac, similar tothe presentation of all stimuli in the toycategory (e.g., a group of cars) and manystimuli in the snack category (a plate of raisins).Another possible reason for Isaac’s success insharing during the presentation of across-category probe stimuli may be the additional


prompting and behavior-specific praise usedduring training. Additional research may showthat these more comprehensive promptingprocedures and praise statements lead to morerobust generalization of sharing.

For Steven, failure to generalize sharing inacross-category probe trials during treatmentmay have occurred because his probe trialinvolved the scooter, which was the only trialto take place in the hallway immediately outsidethe office that served as the experimentalsetting. Thus, his failure to share may havebeen due to an idiosyncrasy of the hallway. Hedid demonstrate across-category generalization,however, during both posttreatment generaliza-tion sessions by offering to share the scooter(probe item) and the basketball (novel item)with a peer.

For prosocial behavior to be maintainedbeyond training conditions, natural contingen-cies of reinforcement must operate on thebehavior. For the specific response of offeringto share, the naturally occurring reinforcers maybe the positive response from, and socialexchange with, the sharing recipient. Offeringto share may not occur in some children withsevere social deficits (e.g., children with autism),because the natural contingency of the responsemay not function as a reinforcer. Thus, theresponse of offering to share often requiresteaching with programmed arbitrary reinforce-ment. We successfully shifted from arbitraryreinforcement to natural contingencies byinitially using tokens (arbitrary reinforcement)paired with a natural verbal response (theexperimenter responding ‘‘thanks’’ or ‘‘sure’’while receiving the offered item), then system-atically fading token reinforcement across twoadditional phases after the criterion level ofoffering to share.

Results indicated that a single presentation ofa video model following response errors hadlimited success in establishing an offer to share.It is possible that the video model was not aseffective in the present study compared to that

used by Reeve et al. (2007) because the videomodel did not match the in vivo exemplar foreach training trial. Although there is noempirical precedent for presenting a videomodel that does not match the presented trial,we believed it was an interesting parameter toinvestigate. A future study might directlyexamine differences in the effects of matchedand nonmatched video models on skill acqui-sition and generalization. It is also possible thata live model may have been more effective thana video model. Additional studies may comparethe effectiveness of live and video models forteaching sharing and other prosocial skills (e.g.,Geiger, LeBlanc, Dillon, & Bates, 2010).

The selection of the stimuli used duringsharing scenarios in the present study was basedon normative observations by the experimenterand reports from the participants’ teachersregarding types of items the participants mightbe expected to share in the classroom. Onelimitation, however, is that formal preferenceassessments were not used to evaluate or selectthe sharing stimuli (Carr, Nicolson, & Higbee,2000; DeLeon & Iwata, 1996; Fisher et al.,1992). More rapid skill acquisition and morerobust generalization and maintenance ofsharing may have been obtained if preferencesfor the stimuli had been determined. Forexample, sharing a highly preferred item maybe aversive, thus requiring additional prompt-ing and more potent contrived reinforcers. Incontrast, sharing a less preferred item mayrequire little training or contrived reinforce-ment. In fact, anecdotal information regardingpreferences was used to select Isaac’s snackcategory. Based on his reported history ofdisplaying tantrums when preferred foods wererestricted, we selected snack foods that werepredicted to be less preferred to reduce thepotential for tantrums. Future studies mightexamine the relation between the preferencelevel of shared stimuli (e.g., highly, moderately,and least preferred) and speed of acquisitionand generalization of sharing.


In addition to differences in preference, thelikelihood of sharing an item may also beaffected by the quantity of stimuli available. Forexample, when children interact with multiplesof the same item, they may be more likely toshare that item because others are still in theirpossession. In contrast, interaction with onlyone item may make it less likely that the itemwill be shared, because sharing would involvetotal removal of the item. Both single items andgroups of items were used in the current study,so experimenters may wish to evaluate thisparameter under controlled conditions in futurestudies.

Additional applications of the proceduresused in this study to other areas of prosocialbehavior may identify more efficient andeffective procedures for establishing generalizedrepertoires of prosocial skills in children withautism. These repertoires would ensure thatchildren with autism have more opportunitiesto interact effectively with others and toexperience greater acceptance by their peers(Bryant & Budd, 1984; Cooke & Apolloni,1976; Eisenberg & Mussen, 1989).

REFERENCES

Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Doesthe autistic child have a ‘‘theory of mind’’? Cognition,21, 37–46.

Barton, E. J., & Ascione, F. R. (1979). Sharing inpreschool children: Facilitation, stimulus generaliza-tion, response generalization, and maintenance.Journal of Applied Behavior Analysis, 12, 417–430.

Bryant, L. E., & Budd, K. S. (1984). Teachingbehaviorally handicapped preschool children to share.Journal of Applied Behavior Analysis, 17, 45–56.

Carr, J. E., Nicolson, A. C., & Higbee, T. S. (2000).Evaluation of a brief multiple-stimulus preferenceassessment in a naturalistic context. Journal of AppliedBehavior Analysis, 33, 353–357.

Chandler, K. L., Lubeck, R., & Fowler, S. A. (1992).Generalization and maintenance of preschool chil-dren’s social skills: A critical review and analysis.Journal of Applied Behavior Analysis, 25, 414–428.

Cooke, T. P., & Apolloni, T. (1976). Developing positivesocial-emotional behaviors: A study of training andgeneralization effects. Journal of Applied BehaviorAnalysis, 9, 65–78.

DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of amultiple-stimulus presentation format for assessingreinforcer preferences. Journal of Applied BehaviorAnalysis, 29, 519–532.

DeQuinzio, J. A., Townsend, D. B., & Poulson, C. L.(2008). The effects of forward chaining and contin-gent social interaction on the acquisition of complexsharing responses by children with autism. Research inAutism Specturm Disorders, 2, 264–275.

Eisenberg, N., & Fabes, R. A. (1998). Prosocial behavior.In W. Damon (Ed.), Handbook of child psychology:Vol. 3. Social, emotional and personality development(5th ed., pp. 701–778). New York: Wiley.

Eisenberg, N., & Mussen, P. H. (1989). The roots ofprosocial behavior in children. Boston: CambridgeUniversity Press.

Engelmann, S., & Carnine, D. (1982). Theory ofinstruction: Principles and applications. New York:Irvington.

Fisher, W., Piazza, C. C., Bowman, L. G., Hagopian, L.P., Owens, J. C., & Slevin, I. (1992). A comparisonof two approaches for identifying reinforcers forpersons with severe and profound disabilities. Journalof Applied Behavior Analysis, 25, 491–498.

Geiger, K. B., LeBlanc, L. A., Dillon, C. M., & Bates, S. L.(2010). An evaluation of preference for video and invivo modeling. Journal of Applied Behavior Analysis, 43,279–283.

Gena, A., Krantz, P. J., McClannahan, L. E., & Poulson,C. L. (1996). Training and generalization of affectivebehavior displayed by youth with autism. Journal ofApplied Behavior Analysis, 29, 291–304.

Horner, R. H., Eberhard, J. M., & Sheehan, M. R.(1986). Teaching generalized table bussing: Theimportance of negative teaching examples. BehaviorModification, 10, 457–471.

Horner, R. H., Sprague, J., & Wilcox, B. (1982). Generalcase programming for community activities. In B.Wilcox & G. T. Bellamy (Eds.), Design of high schoolprograms for severely handicapped students (pp. 61–98).Baltimore: Brookes.

Krantz, P. J., MacDuff, G. S., Wadstrom, O., &McClannahan, L. E. (1991). Using video withdevelopmentally disabled learners. In P. W. Dowrick(Ed.), A practical guide to using video in the behavioral

sciences (pp. 256–266). New York: Wiley.LeBlanc, L. A., Coates, A. M., Daneshvar, S., Charlop-

Christy, M. H., Morris, C., & Lancaster, B. M.(2003). Using video modeling and reinforcement toteach perspective-taking skills to children with autism.Journal of Applied Behavior Analysis, 36, 253–257.

Nikopoulos, C. K., & Keenan, M. (2004). Effects of videomodeling on social initiations by children withautism. Journal of Applied Behavior Analysis, 37,93–96.

Reeve, S. A., Reeve, K. F., Townsend, D. B., & Poulson,C. L. (2007). Establishing a generalized repertoire ofhelping behavior in children with autism. Journal ofApplied Behavior Analysis, 40, 123–136.


Rheingold, H. L., & Hay, D. F. (1980). Prosocialbehavior of the very young. In G. S. Stent (Ed.),Morality as a biological phenomenon (pp. 93–108).Berkeley: University of California Press.

Rogers-Warren, A., & Baer, D. M. (1976). Correspon-dence between saying and doing: Teaching childrento share and praise. Journal of Applied BehaviorAnalysis, 9, 335–354.

Rutter, M. (1978). Diagnosis and definition of childhoodautism. Journal of Autism and Childhood Schizophre-nia, 8, 139–161.

Sawyer, L. M., Luiselli, J. K., Ricciardi, J. N., & Gower,J. L. (2005). Teaching a child with autism toshare among peers in an integrated preschoolclassroom: Acquisition, maintenance and socialvalidation. Education and Treatment of Children, 28,1–10.

Schrandt, J. A., Townsend, D. B., & Poulson, C. L.(2009). Teaching empathy skills to children withautism. Journal of Applied Behavior Analysis, 42,17–32.

Sprague, J. R., & Horner, R. H. (1984). The effects ofsingle instance, multiple instance and general casetraining on generalized vending machine use bymoderately and severely handicapped students. Jour-nal of Applied Behavior Analysis, 17, 273–278.

Stokes, T. F., & Baer, D. M. (1977). An implicittechnology of generalization. Journal of AppliedBehavior Analysis, 10, 349–367.

Stokes, T. F., & Osnes, P. G. (1989). An operant pursuitof generalization. Behavior Therapy, 20, 337–355.

Taylor, B. A., Levin, L., & Jasper, S. (1999). Increasingplay-related statements in children with autismtoward their siblings: Effects of video modeling.Journal of Developmental and Physical Disabilities, 11,253–264.

Volkmar, F. R., Carter, A., Sparrow, S. S., & Cicchetti, D.V. (1993). Quantifying social development in autism.Journal of the American Academy of Child andAdolescent Psychiatry, 32, 627–632.

Warren, S. F., Rogers-Warren, A., & Baer, D. M. (1976).The role of offer rates in controlling sharing by youngchildren. Journal of Applied Behavior Analysis, 9,491–497.

Wing, L. (1988). The continuum of autistic characteris-tics. In E. Schopler & G. B. Mesibov (Eds.), Diagnosisand assessment of autism (pp. 91–110). New York:Plenum.

Received July 8, 2009Final acceptance July 27, 2010Action Editor, Linda LeBlanc


Documents

USING MULTIPLE-EXEMPLAR TRAINING TO TEACH A … · 2013. 8. 2. · Video Models The video models used to teach sharing depicted two peers sharing an activity shown from a third-person