ERIC EJ931955: Using Multiple-Exemplar Training to Teach a Generalized Repertoire of Sharing to Children with Autism

2011·0.54 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2011, 44, 279–294 NUMBER2 (SUMMER2011)
USING MULTIPLE-EXEMPLAR TRAINING TO TEACH A
GENERALIZED REPERTOIRE OF SHARING TO CHILDREN
WITH AUTISM
DENISE MARZULLO-KERTH, SHARON A. REEVE, AND KENNETH F. REEVE
CALDWELLCOLLEGE
AND
DAWN B. TOWNSEND
INSTITUTEFOREDUCATIONALACHIEVEMENT
The current study examined the utility of multiple-exemplar training to teach children with
autismtoshare.Stimulifrom3of4categoriesweretrainedusingatreatmentpackageofvideo
modeling,prompting,andreinforcement.Offerstoshareincreasedforall3childrenfollowing
theintroductionoftreatment,withevidenceofskillmaintenance.Inaddition,within-stimulus-
category generalization of sharing was evident for all participants, although only 1 participant
demonstrated across-category generalization of sharing. Offers to share occurred in a novel
setting, with familiar and novel stimuli, and in the presence of novel adults and peers for all
participants during posttreatment probes.
Keywords: autism,generalization,multiple-exemplartraining,socialbehavior,sharing,video
modeling
_______________________________________________________________________________
Prosocial behavior refers to a broad class of ties for positive social interactions with peers
responses that includes smiling, cooperating, andisanintegralcomponentofinteractiveplay.
taking turns, making friends, expressing empa- Severalstudiesillustratesuccessfulprocedures
thy, helping others, and sharing (Barton & for teaching children of typical development
Ascione, 1979; Bryant & Budd, 1984; Chan- how to share with their peers. Warren et al.
dler, Lubeck, & Fowler, 1992; Cooke & (1976) for example, used modeling and rein-
Apolloni, 1976; Reeve, Reeve, Townsend, & forcement to increase offers to share and
Poulson, 2007; Rogers-Warren & Baer, 1976; corresponding acceptances of share offers in
Warren,Rogers-Warren,&Baer,1976).Bryant two groups of preschool children with reported
andBudd(1984)suggestedthatsharingmaybe generalization to a new setting. Barton and
an especially important social behavior for Ascione(1979)examinedthegeneralizationand
young children because it increases opportuni- durability of sharing by teaching preschool
children to share verbally (e.g., ‘‘here you
go’’), physically (e.g., handing a toy), or both.
This study was conducted as partial fulfillment of the Physical sharing increased for all children
requirements for the first author’s MA degree in applied during training, but it generalized and main-
behavioranalysisfromCaldwellCollege.Portionsofthese
tained only for children who were taught to
datawerealsopresentedatthe34thannualconventionof
the Association for Behavior Analysis, May, 2008, in share verbally. Verbal sharing increased during
Chicago.WethankthestudentsandstaffattheY.A.L.E. training only for children who were taught to
Schoolfor their participation.
share verbally or both verbally and physically;
Correspondence concerning this article should be
addressed to Sharon A. Reeve, Department of Applied however, verbal sharing did not generalize
Behavior Analysis, Caldwell College, 120 Bloomfield across settings. Bryant and Budd (1984)
Avenue, Caldwell, New Jersey 07006 (e-mail: sreeve@
extended this training package to six preschool-
caldwell.edu).
doi:10.1901/jaba.2011.44-279 ers with disabilities, focusing on teaching
279
280 DENISE MARZULLO-KERTH et al.
specificcomponentsofsharingresponses:offers, children with autism, multiple-exemplar train-
requests,andacceptances.Allthreecomponents ing has been used to teach helping behavior
of sharing increased after implementation of (Reeve et al., 2007), empathy (Schrandt,
instructions, modeling, and behavior rehearsals. Townsend, & Poulson, 2009), and appropriate
Compared to children of typical develop- affect (Gena, Krantz, McClannahan, & Poul-
ment, children with autism often display son, 1996), among others.
especially marked impairments in sharing Stimulithatadequatelyreflectthediversityof
(Baron-Cohen,Leslie,&Frith,1985;Eisenberg the stimulus characteristics likely to be present
& Fabes, 1998; Rheingold & Hay, 1980; under generalization conditions are used during
Rutter, 1978; Volkmar, Carter, Sparrow, & multiple-exemplar training. The identification
Cicchetti, 1993; Wing, 1988). Despite these of these stimuli by the teacher is known as a
deficits, few studies have examined methods to general case analysis (Engelmann & Carnine,
teach children with autism to share. Sawyer, 1982; Horner, Sprague, & Wilcox, 1982;
Luiselli, Ricciardi, and Gower (2005) increased Sprague & Horner, 1984). When responding
verbal and physical sharing in one child with comesunderthecontrolofthestimulusfeatures
autism through the use of priming before play present in the exemplars used during training,
sessions, along with in-session prompts and the presentation of novel stimuli that share
reinforcement;however,theydidnotassessskill these features should also occasion the same
generalization. Recently, DeQuinzio, Town- response.
send, and Poulson (2008) used a forward The purpose of the present study was to
chaining procedure to teach children with extend the findings of DeQuinzio et al. (2008)
autism to approach a peer and emit a show- by establishing a generalized repertoire of
give-play sharing response chain with multiple sharing in four children with autism. A
toy exemplars. Sharing of toys increased for all multiple-exemplar teaching procedure similar
participants and generalized to nontrained toys, to that of Reeve et al. (2007) was used.
novel locations, and different peers. However, Specifically, children were taught to share items
DeQuinzioetal.conductedtrainingandtesting from multiple classesof materials(artmaterials,
with only one class of materials (toys), so it is snack foods, toys, and gym materials), and
not clear whether sharing could be successfully generalization was assessed both within classes
taught using other classes of materials or of materials and across nontaught classes of
whether sharing would generalize to untrained materials. In addition, teaching was conducted
classes of materials. in different settings to promote generalization
Because the utility of a behavior-change acrosslocations,anddiscriminationofnonshar-
procedure is increased when responding gener- ingsituationswastargetedandassessed.Finally,
alizes across persons, settings, and materials we assessed whether sharing was under appro-
(Stokes & Baer, 1977; Stokes & Osnes, 1989), priate stimulus control by sampling situations
it is important to implement procedures that and stimuli for when sharing was appropriate
promote generalization. One strategy to accom- and situations in which it was not. Also similar
plish this involves the teaching of multiple to Reeve et al., a combination of video
exemplars. Multiple-exemplar training has been modeling and prompting was used for error
used to teach numerous skills to individuals correction. Video modeling provided a means
with moderate to severe disabilities, including to standardize the presentation of the models
appropriate vending machine use (Sprague & andhasbeenshowntopromoteacquisitionand
Horner, 1984) and setting and clearing tables generalization of skills for children with autism
(Horner, Eberhard, & Sheehan, 1986). For (Krantz, MacDuff, Wadstrom, & McClanna-
TEACHING SHARING 281
han, 1991; LeBlanc et al., 2003; Nikopoulos & ing machine. Pre- and posttreatment sessions
Keenan, 2004; Taylor, Levin, & Jasper, 1999). were conducted in the usual daily classroom to
assess generalization of sharing.
Materials and target responses for the study
METHOD
were selected based on data collected with first-
Participants
through third-grade general and special educa-
Four children with a diagnosis of autism
tion students during a cumulative 2 hr of
participated. They attended a private school for
observations in public elementary schools
individuals with autism and had received their
conductedduringfreeplay,snacktime,outdoor
diagnoses through independent agencies prior
recess, and gym. Materials used included
to school enrollment. Direct observation and
modeling clay, crayons, stamps, jump rope,
parentorteacherreportsconductedpriortothe
balls, puppets, pretend food, blocks, small
start of the study indicated that none of the
plastic animals, books, cars, and snacks. The
participantsengagedinsharingintheclassroom
majority of the observed sharing statements
or at home. Steven was 8 years 1 month old,
were requests or directions to share another
Isaac was 7 years 10 months old, Bobby was
student’s materials (e.g., ‘‘want one?,’’ ‘‘here
8 years 1 month old, and Aiden was 7 years
you go,’’ ‘‘wanna play?,’’ ‘‘your turn,’’ and
6 months old. Each participant had previous
‘‘here’’).Theclassroomteacherswerealsoasked
experience learning skills with discrete-trial
what types of materials the participants were
instruction and using token-based motivational
likely to use during free time and group time,
systems. According to reports by their teachers,
when offering to share would be appropriate.
each participant also demonstrated the prereq-
The four stimulus categories were art mate-
uisite verbal skills needed to make the target
rials, snack foods, toys, and gym materials.
vocal responses used in the present study (e.g.,
Stimulus categories and items assigned to each
‘‘Want to try it?’’ ‘‘You try it’’), displayed
participant are listed in Table 1. In each of
correct imitation of various verbal and motor
these four categories, five different materials
responses using both in vivo and video models,
served as multiple exemplars of the same
and followed simple directions to engage in
stimulus category. For example, art materials
specified activities.
included crayons, dot paint, markers, colored
Setting and Materials pencils,andglitterpens.SnackfoodsforSteven,
Experimental sessions were conducted in a Bobby, and Aiden included pretzels, cookies,
school office that was not used for the chips, candy, and marshmallows. Specific snack
participants’ daily education. The office con- foods were different for Isaac because he
tained a desk and three chairs, bookshelf, filing displayed disruptive behavior in response to
cabinet, computer, printer, and telephone. the removal of the original snacks (which were
Additional materials included all relevant stim- highly preferred) during the initial treatment
ulus materials (i.e., toys, gym equipment, art session. To reduce disruption, snack foods for
materials, and snacks), a portable DVD player, Isaac included carrots, celery, apples, and
small audio voice recorders, a token board, a raisins. During pre- and postintervention gen-
video camera with tripod, and data sheets. eralization probes, however, snack stimuli for
Approximately every 2 weeks, experimental Isaac were the same as those used for the other
sessions were conducted in the school’s kitchen three participants. Toys included Lego blocks,
to promote generalization. The kitchen con- cars, magnets, Play-Doh, and Peg-Board. Gym
tainedtworoundtableswith fivechairsateach, materials included a Velcro mitt and ball,
stove, sink, refrigerator, dishwasher, and vend- basketball, scooter, Hippity-Hop, and Velcro
282 DENISE MARZULLO-KERTH et al.
Table 1
Assignment of StimuliandVerbal Responses AcrossParticipants
Steven Isaac Bobby Aiden
Teaching Art: Snack: Toys: Gym:
crayons celery magnets Hippity-Hop
dotpaint apples Play-Doh Velcromittandball
markers raisins Legoblocks magneticdarts
Snack: Toys: Gym: Art:
pretzels cars scooter crayons
cookies magnets Hippity-Hop coloredpencils
chips Play-Doh Velcromittandall dotpaint
Toys: Gym: Art: Snack:
Legoblocks basketball markers candy
cars scooter coloredpencils pretzels
magnets Velcromittandball crayons cookies
Within-category Art:coloredpencils Snack:carrots Art:dotpaint Art:markers
probe Snack:candy Toys:Legoblocks Gym:basketball Snack:chips
Toys:Play-Doh Gym:Hippity-Hop Toys:cars Gym:scooter
Across-category Gym:scooter Art:crayons Snack:chips Toys:cars
probe
Verbalresponses ‘‘Wouldyouliketotrythis?’’ ‘‘Trythis.’’ ‘‘Whydon’tyoutry?’’ ‘‘Here,youtryit.’’
‘‘Doyouwanttotry?’’ ‘‘Doyouwanttotry?’’ ‘‘Trythis.’’ ‘‘Wouldyouliketotrythis?’’
‘‘Here,youtryit.’’ ‘‘Here,youtryit.’’ ‘‘Doyouwanttotry?’’ ‘‘Doyouwanttotry?’’
‘‘Trythis.’’
Pre-and Snack: Snack: Snack: Snack:
posttreatment marshmallows(G) marshmallows(G) chips(P) marshmallows(G)
generalization candy(P) candy(P) marshmallows(G) chips(P)
trialsa cookies(T) pretzels(T) pretzels(G) candy(T)
Art: Art: Art: Art:
glitterpens(G) crayons(P) glitterpens(G) glitterpens(G)
coloredpencils(P) glitterpens(G) dotpaint(P) markers(P)
markers(T) markers(G) crayons(T) dotpaint(T)
Gym: Gym: Gym: Gym:
scooter(P) Velcrodarts(G) Velcrodarts(G) Velcrodarts(G)
basketball(G) Hippity-Hop(P) basketball(P) magneticdarts(P)
Velcrodarts(G) Velcromittandball(T) Hippity-Hop(T) Hippity-Hop(T)
Toys: Toys: Toys: Toys:
Peg-Board(G) Peg-Board(G) Peg-Board(G) cars(P)
Play-Doh(P) Legoblocks(P) cars(P) Peg-Board(G)
Legoblocks(T) cars(T) Play-Doh(T) magnets(G)
aIncludesstimuliusedonlyduringpre-andposttreatmentgeneralizationtrials(G),probetrialsduringexperimental
sessions (P), andteaching sessions (T).
darts. One exemplar in each category was natural settings, no assessment was made
randomly selected for use in pre- and posttreat- regarding the participants’ relative preferences
ment generalization probes only (i.e., glitter for any of the materials.
pens, marshmallows, peg board, and Velcro
darts). In addition, materials that served as Video Models
discriminative stimuli for nonsharing responses The video models used to teach sharing
were present, including papers to put in a depicted two peers sharing an activity shown
backpack,anacademicworksheet,booksonthe from a third-person viewpoint (i.e., scene
floorthatneededtobeplacedonatable,adirty perspective). Specifically, a 7-year-old boy was
tabletop and a towel, and an article of clothing. first shown either sitting at a table or standing,
It should be noted that although the materials engaged in an activity. A second 7-year-old boy
used for sharing and nonsharing trials were all then came on camera, approached the first boy,
selected based on observations of peers in and stood within 0.6 m of him. Following the
TEACHING SHARING 283
Table 2
Video Model Types AcrossParticipants
Videomodel
Verbalresponse Stimulussets
Steven ‘‘Wouldyouliketotrythis?’’ cookies,cars
‘‘Doyouwanttotry?’’ markers,crayons
‘‘Trythis.’’ magnets,chips
Isaac ‘‘Trythis.’’ magnets,chips
‘‘Doyouwanttotry?’’ Velcromittandball,pretzel
‘‘Here,youtryit.’’ ball,Play-Doh
Bobby ‘‘Whydon’tyoutry?’’ coloredpencils,Velcromittandball
‘‘Trythis.’’ Play-Doh,scooter
‘‘Doyouwanttotry?’’ magnets,markers
Aiden ‘‘Here,youtryit.’’ football,coloredpencils
‘‘Wouldyouliketotrythis?’’ cookies,cars
‘‘Doyouwanttotry?’’ pretzels,Velcromittandball
approach of the second boy, the first boy held correct nonsharing response was scored if the
outhis item and emitteda verbal offer to share. child responded on a nonsharing trial by
Sixvideomodelswereusedforeachparticipant, completing the task without offering to share,
and each video model was approximately 8 to eitherphysicallyorverbally.Table 2depictsthe
10 s in duration. Stimuli featured in each video randomized assignment of verbal responses
model corresponded to stimuli used during taught to participants.
training trials for that particular participant. A teacher with master’s level training in
Similarly, verbal offers to share emitted by the behavior analysis served as a second observer
peer in the video model were verbal responses and scored sharing and nonsharing either in
assigned to that particular participant during vivo or via videotapes of a session. For each
treatment. Table 2 provides a list of the stimuli trial, the scoring of the two observers was
and responses featured in the video models for compared, and only an exact match was
each participant. For example, one of Steven’s considered an agreement. Percentage agreement
video models depicted a peer coloring with was calculated by dividing the number of
crayons who asked, ‘‘Do you want to try?’’ in agreements by the number of agreements plus
response to the approach of the second peer. disagreements and multiplying by 100%.
Interobserver agreement data were collected on
Response Measurement the dependent variable during 54% of Steven’s
Trainedobservers scoredparticipantrespons- sessions, 44% of Isaac’s sessions, 47% of
es on a trial-by-trial basis using pencils and Bobby’s sessions, and 40% of Aiden’s sessions.
paper data sheets. A correct offer to share was The mean percentage of agreement across all
scored only if both the motor and verbal trial types (training, probe, and nonsharing
component occurred. A correct motor response trials) and across all participants was 99%
consistedoftheparticipantholdingoutanitem (session range, 97% to 100%). A trained
to the experimenter within 5 s of her approach observer also collected data on the accuracy of
to within 0.6 m of the participant. A correct the experimenter’s presentation of discrimina-
verbal response consisted of a request for the tive stimuli, reinforcement, trial types, and
experimenter to engage in an activity with the video model error correction. These procedural
presented item (e.g., ‘‘Do you want to try?’’; integrity data were collected during 95% of all
‘‘Here, you try it’’) or an approximation of the baseline, treatment, maintenance, and pre- and
response (e.g., ‘‘Do you want it?’’; ‘‘try’’). A posttreatment sessions. The mean percentage
284 DENISE MARZULLO-KERTH et al.
accuracy on procedural components was 99% promote response generalization, three of the
(session range, 98% to 100%). In addition, a five experimenter-defined verbal offers to share
second observer scored procedural integrity for were randomly assigned to each participant and
46% of Steven’s sessions, 34% of Isaac’s were rotated across trials. At the start of each
sessions, 36% of Bobby’s sessions, and 34% of trial, the experimenter approached and stood
Aiden’s sessions. The mean percentage of within 0.6 m of the participant while he was
agreement on the accurate presentation of engaged in a specific activity. The purpose of
procedural components was 99% (session presenting these discriminative stimuli was to
range, 98% to 100%). set the occasion for the participant to offer to
share the item or items with which he was
Design and Assignment of Categories
engaged. Finally, five nonsharing discrimina-
A concurrent multiple-probe design across
tion trials were interspersed among the nine
participantswasused.Threeofthefourpossible
teaching and four probe trials. To reduce
stimuluscategorieswereassignedtothetraining
potential order or sequence effects, the order
condition for each participant; the fourth
ofthetrialswasrandomized,withtheexception
category was used to assess across-category
thatallsessionsbeganandendedwithatraining
generalization of sharing responses (see Ta-
trial.
ble 1). Each category contained five possible
Baseline.Theparticipantwasseatedatadesk.
stimulus exemplars: Three were directly target-
The experimenter presented stimulus materials
ed as stimulus exemplars (partially counterbal-
for that trial (e.g., crayons and a blank coloring
anced across participants), a fourth exemplar
page) with a verbal direction to engage in the
was used to assess within-category generaliza-
targetactivity(e.g.,‘‘colorthepicture’’).Within
tionof sharing during trainingsessions, andthe
5sofpresentingthematerials,theexperimenter
fifthexemplarwasusedtofurtherassess within-
approached the participant and stood within
and across-category generalization during the
0.6 m of him. If the participant emitted a
pre-andpostinterventiongeneralizationprobes.
correct verbal and motor offer to share, the
For example, Steven’s teaching trials included
experimenter responded ‘‘sure’’ or ‘‘thanks,’’
crayons, dot paint, and markers (art materials
took the offered item, and manipulated the
category); pretzels, cookies, and chips (snack
itemappropriately.Nopromptsormodelswere
foods category); and Lego blocks, cars, and
provided, and no tokens were delivered for
magnets (toy category). Probe trials for Steven
sharing.Theexperimenterdeliveredtokensona
included a scooter (gym materials category) to
variable-ratio (VR) 2 schedule for following the
assess across-category generalization, colored
direction to engage in the task and attending to
pencils to assess within-category generalization
the experimenter. The experimenter did not
forartmaterials,candytoassesswithin-category
respondiftheparticipantemittedonlyacorrect
generalizationforsnackfoods,andPlay-Dohto
motor response or a correct verbal response. If
assess within-category generalization for toys.
the participant did not emit a correct response
Procedure within 5 s of the presentation of the discrim-
Eachbaselineandtreatmentsessionconsisted inative stimuli, the experimenter removed the
of 18 trials. Of these, nine were teaching trials materials and walkedaway. After a 5-s intertrial
(threeexemplarsfromeachofthethreetraining interval, the experimenter presented the partic-
categories). Four additional trials were general- ipant with the next activity and a new trial
ization probes with the fourth exemplar from began. At the conclusion of a session, partici-
each of the three training categories and one pants could trade in their tokens for previously
exemplarfromthefourthstimuluscategory.To selected activities or materials (e.g., access to a
TEACHING SHARING 285
video game, preferred snacks or toys) that were for that trial a second time and approached and
not used during the sessions. stood within 0.6 m of the participant. If the
Treatment. The steps of the procedure are participant offered to share within 5 s, the
depicted in Figure 1. As in baseline, the experimenter accepted the offered item; re-
experimenter presented the participant with sponded with ‘‘yes,’’ ‘‘sure,’’ or ‘‘thank you’’;
stimulus materials, gave the direction to engage manipulated the item appropriately; and deliv-
inthetargetactivity,waited5s,approachedthe ered a token to the participant’s token board. If
participant, and stood within 0.6 m of him. the participant did not offer to share within 5 s
Following a correct sharing response during of the second opportunity to do so, the
training trials, the experimenter accepted the experimenter provided a physical prompt by
offered item; confirmed the offer to share by using hand-over-hand guidance to assist the
saying ‘‘yes,’’ ‘‘sure,’’ or ‘‘thank you’’; manip- child in handing the items with which he was
ulated the offered item appropriately; and then engaged to the experimenter. Simultaneously,
delivered a token. After appropriately manipu- the experimenter activated a voice recorder that
lating the item and returning it to the emittedarecordingofthetargetvocalresponse.
participant, the experimenter walked away and If the participant emitted only one component
gave the direction to clean up. Materials were of the response (e.g., physical or vocal), the
removed from the table and the trial ended. A experimenter delivered the corresponding
new trial began following a 5-s intertrial promptfortheabsentcomponent.Forexample,
interval. if the child handed an item to the experimenter
If the participant did not respond within 5 s but did not emit a target vocal response, the
of presentation of the discriminative stimuli, experimenter provided only an audio prompt
responded incorrectly, or emitted only one by activating the voice recorder. If the child
component of the sharing response (e.g., motor emitted a target vocal response but did not
response or vocal response alone), the experi- physically offer to share, the experimenter
menterremovedthematerialsandimplemented provided only a physical prompt with hand-
an error-correction procedure similar to that over-hand guidance. The physical and auditory
described by Reeve et al. (2007). First, the prompts were repeated until the participant
experimenter presented a video model of a emitted the correct response independently.
scenario in which appropriate sharing was As noted previously, generalization probe
depicted. Video models were used only when trials and nonsharing trials were interspersed
a participant failed to emit appropriate sharing with training trials. During probe trials, the
behavior during training trials. For ease of procedure was identical to that during baseline.
implementation and to promote generalization During nonsharing trials, the experimenter
of sharing, the video models did not necessarily presented the nonsharing stimuli to the partic-
show the same activity in which the participant ipants and emitted an instruction appropriate
was engaged during that trial. For example, a for those stimuli (e.g., ‘‘Can you please put the
videomodeldepictingapeereatingcookiesand books on the shelf?’’; ‘‘Can you please wipe the
asking ‘‘Would you like to try this?’’ could be table?’’; ‘‘Put the papers in your backpack’’).
presented to Steven during a trial in which he Whentheparticipantcompletedthenonsharing
was playing with cars and the target response task (e.g., placed papers in a backpack), the
was ‘‘Would you like to try this?’’ experimenter delivered a token and the trial
None of the participants required prompting ended. If a child emitted a sharing response
to watch the video. When the video model (e.g.,‘‘Wanttotry?’’)duringanonsharingtrial,
ended,theexperimenterpresentedthematerials the experimenter paused for 5 s and then
286 DENISE MARZULLO-KERTH et al.
Figure 1. Instructional procedure for sharing trialsduring treatment.
TEACHING SHARING 287
presented the same nonsharing trial a second ule, the posttreatment generalization probe
time while verbally instructing the child to trials were conducted. The schedule of rein-
complete the task. This continued until the forcement for Aiden was not systematically
child completed the task and did not emit a thinned due to his completion of participation
sharing response. These nonsharing trials were at the end of the school year.
included to ensure that the participants attend- Pre- and posttreatment generalization probes.
ed to therelevant discriminative stimuli that set The experimenter conducted probes assessing
the occasion to offer to share and to prevent generalizationofsharingtonovelsettings,novel
overgeneralization of the target response to people, and novel stimuli during two sessions
nonsharing stimuli. prior to baseline and two sessions after the
One participant (Isaac) engaged in disruptive trainingcriterionwasmet.The12trialsineach
behavior during the general prompting proce- session consisted of three trials of teaching
dure during training trials. To facilitate skill stimuli, four trials of probe stimuli that served
acquisition and reduce the occurrence of as within- and across-category generalization
disruption, additional prompts were used probes during training sessions, three trials of
during treatment. A gestural prompt was added the fifth stimuli from each of the three training
to the discriminative stimuli in each trial, and a categories that were not presented during
behavior-specific praise statement was delivered training, and two trials of additional stimuli
for offers to share. During teaching trials, after fromthefourth,nontrainingcategory.Withthe
presenting the stimulus materials and giving exception of the scooter, Hippity-Hop, dot
Isaacaverbaldirectiontoengageintheactivity, paint,andmagnets,stimuliusedinthepre-and
the experimenter approached Isaac, stood posttreatment generalization probes were never
within 0.6 m, and extended an open palm presented in experimental sessions. Table 1
toward him. Following errors, the video model depicts stimulus types used in pre- and
andauditory or manualprompts were delivered posttreatment generalization probe trials for
as in the treatment procedure. In addition, each participant. During these probe sessions, a
when Isaac shared correctly, the experimenter novel instructor presented the stimulus materi-
emitted a behavior-specific praise statement alsandgavethedirectiontoengageinthetarget
(e.g., ‘‘Great! You handed it to me and said, activity. In addition, 5 s after the presentation
‘try this.’’’). The experimenter no longer of the materials, a peer was directed to ‘‘see
delivered behavior-specific praise when Isaac what [participant’s name] is doing’’ and to
met the mastery criterion and no longer approachtheparticipantandstandwithin0.6m
delivered the gestural prompt after the schedule of him. The peer, who was approximately the
of reinforcement was thinned. same age as the participant, was another child
The mastery criterion was offering to share with autism. This trial format was unique
independentlyonatleast89%(eightofnine)of because it was a peer who approached the
the teaching trials for four consecutive sessions. participant rather than the person who initially
The fixed-ratio (FR) 1 schedule of token presentedthematerials.Theparticipantdidnot
delivery for correct responding was thinned to receive tokens for correct responses or on-task
VR 2 when the participant met the mastery behavior, nor did he receive prompts or video
criterion. After criterion-level responding oc- models following incorrect responses. No
curredduringteachingtrialsfortwoconsecutive additional prompts were provided if the
sessions at VR 2, the schedule was thinned to participant emitted a correct offer to share,
VR 5. After two consecutive sessions of and the peer accepted the item. If the
criterion-level responding on the VR 5 sched- participant emitted a correct response and the
288 DENISE MARZULLO-KERTH et al.
peer did not take the item, the instructor from 0% during baseline to 75% during
directed the peer to accept the item and to treatment. Bobby demonstrated an increase in
manipulate it appropriately (e.g., ‘‘You can eat offers to share on teaching trials from 0%
the chip,’’ or ‘‘Take the crayon and color with during baseline to 100% during treatment and
Steven.’’). The materials were removed and the from 0% during baseline to 50% during
trial ended if the participant did not emit a treatment for probe trials. Aiden also demon-
correct response within 5 s of the presentation stratedanincreaseinofferstoshareonteaching
of the discriminative stimuli. trials from a mean of 2% during baseline to
Maintenance. Maintenance sessions were 89% during treatment, and from 0% during
conducted 1, 2, and 5 weeks after the baseline to 75% during treatment for probe
posttreatment generalization probes for Steven, trials.
and 1, 2, and 3 weeks after the posttreatment Higher percentages of correct responses
generalization probes for Isaac and Bobby. occurred during within-category probe trials
Maintenance sessions were not conducted for (Steven, M 5 80%; Isaac, M 5 67%; Bobby,
Aidenbecausetheschoolyearendedandhewas M 5 63%; Aiden, M 5 60%) than during
no longer available to participate. Procedures
across-category probe trials (data not shown).
were identical to those in baseline except that
Specifically, Steven offered to share during
the experimenter did not deliver tokens for on-
across-category generalization probe trials dur-
task behavior.
ing only one treatment session (M 5 9%).
Bobby never demonstrated across-category gen-
RESULTS eralization during treatment, and Aiden dem-
onstrated generalization during two treatment
Figure 2 shows the percentage of teaching,
sessions (M 5 13%). Isaac was the only
probe, and nonsharing trials in which each
participant who demonstrated across-category
participant independently offered to share,
generalizationonseveraloccasions(M556%).
across consecutive sessions. Probe data are
Throughout experimental sessions, all partic-
collapsed across both within-category and
ipants displayed some degree of vocal response
across-category generalization trials. During
generalization beyond the three scripted target
baseline, none of the participants offered to
phrases they were taught. Steven made nine
share during any trial, with the exception of
Aiden,whohandedanonpreferredfooditemto unscripted vocal offers to share, Isaac made
the experimenter on two occasions and said seven,Bobbymadetwo,andAidenmadeseven.
‘‘here.’’Followingthesuccessiveintroductionof Examples of nontaught offers to share included
treatment across participants, systematic in- ‘‘Wouldyouliketodraw?’’;‘‘Wouldyouliketo
creases in offering to share occurred during build?’’; ‘‘ride the scooter’’; ‘‘try, please.’’
both teaching and probe trials. The percentage None of the participants offered to share
of trials in which Steven offered to share during during any trial presented during the two
teachingtrials systematically increased from 0% pretreatment generalization probe sessions (Fig-
during baseline to 100% during treatment. ure 2). After each participant met the mastery
During probe trials, the percentage of trials in criterion, however, the percentage of trials in
which he offered to share increased from 0% which each participant offered to share in the
during baseline to 75% during treatment. presenceofbothfamiliarandnovelstimuli,ina
Isaac’s offers to share during teaching trials novel setting, and with novel adults and peers
changed in a similar manner, increasing from increased compared to the pretreatment gener-
0% during baseline to 100% during treatment. alization probes. Steven offered to share with a
His offers to share during probe trials increased peer on 100% of the trials presented for both