ERIC EJ832373: Teaching Empathy Skills to Children with Autism

2009·0.31 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2009, 42, 17–32 NUMBER1 (SPRING2009)
TEACHING EMPATHY SKILLS TO CHILDREN WITH AUTISM
JESSICA A. SCHRANDT
QUEENSCOLLEGEANDTHEGRADUATECENTEROFTHECITY
UNIVERSITYOFNEWYORK
DAWN BUFFINGTON TOWNSEND
INSTITUTEFOREDUCATIONALACHIEVEMENT
AND
CLAIRE L. POULSON
QUEENSCOLLEGEANDTHEGRADUATECENTEROFTHECITY
UNIVERSITYOFNEWYORK
The purpose of this study was to teach empathetic responding to 4 children with autism.
Instructorspresentedvignetteswithdollsandpuppetsdemonstratingvarioustypesofaffectand
used prompt delay, modeling, manual prompts, behavioral rehearsals, and reinforcement to
teach participants to perform empathy responses. Increases in empathetic responding occurred
systematicallywiththeintroductionoftreatmentacrossallparticipantsandresponsecategories.
Furthermore, responding generalized from training to nontraining probe stimuli for all
participants.Generalizationoccurredfromdollsandpuppetstoactualpeopleinanontraining
setting for 2 participants. Generalization was observed initially to the nontraining people and
setting for the other participants, but responding subsequently decreased to baseline levels.
Introduction oftreatment inthis setting produced rapidacquisition of targetskills.
DESCRIPTORS: autism,empathy, socialskills
_______________________________________________________________________________
Deficits in empathy and perspective taking states, take the perspective of another person,
have been well documented in children with and respond with empathy.
autism (e.g., Rutter, 1978; Sigman, Kasari, Empathy and perspective taking serve an
Kwon, & Yirmiya, 1992; Yirmiya, Sigman, important role in what Rheingold and Hay
Kasari,&Mundy,1992).Forexample,Yirmiya (1980) called prosocial behavior (e.g., helping,
et al. demonstrated that after watching video sharing, turn taking). Eisenberg (1992) and
segments of children experiencing different Rheingold, Hay, and West (1976) observed
events and emotions, high-functioning children prosocial behavior, such as sharing and giving
with autism were less able than their typically comfort, in typically developing children as
developing peers to label others’ emotional young as 18 months old. Typically developing
childrendisplayperspective-takingskillsaround
4 years of age (MacNamara, Baker, & Olson,
1976; Wimmer & Perner, 1983). It is easy to
We thank Alana Bellizzi, Casey Gaetano, Julia Smith,
see why such behavior is important in the
and Jennifer Williams (teachers at the Institute for
Educational Achievement), and Paul Argott, all of whom development of social relationships. Peers and
assisted in carrying out this study. Special thanks to the family members may initiate social interaction
participants and their families for making this study
more frequently with children who take the
possible.
Correspondence concerning this article should be perspective of and demonstrate empathy to-
addressed to Jessica Schrandt, 4048 Harriet Ave., wards others.
Minneapolis, Minnesota 55409 (e-mail: jessyschrandt@
Althoughmanyresearchershavedocumented
yahoo.com).
doi:10.1901/jaba.2009.42-17 deficits in empathy andperspective-taking skills
17
18 JESSICA A. SCHRANDT et al.
in individuals with autism (e.g., Baron-Cohen, limited number of cues when presented with
Leslie, & Frith, 1985; Bauminger & Kasari, complex stimuli (Bailey, 1981; Lovaas, Koegel,
1999; Sigman et al., 1992; Yirmiya et al., & Schreibman, 1979; Schreibman, Koegel, &
1992), fewer have focused on the extent to Craig, 1977). When learning new discrimina-
which these and related social skills can be tions,individualswithautismmayattendtofew
taught. In the studies reviewed, modeling, or irrelevant stimuli in the environment. Thus,
prompting, and reinforcement were the most failures in empathy (or other complex social
frequently used and effective procedures for behavior) may not reflect a deficit of necessary
teaching social skills. responses in the repertoire but rather an
Gena, Krantz, McClannahan, and Poulson inability to differentiate the stimuli in the
(1996) used modeling, prompting, and rein- presence of which specific responses (e.g.,
forcement to teach 4 children with autism offering assistance, demonstrating interest)
contextually appropriate affective responding would be appropriate.
to various scenarios (e.g., showing sympathy, In the natural environment, discriminative
laughing about absurdities). The treatment stimuli for empathetic responding are both
resulted in increased appropriate affect across varied and complex. In previous studies,
participants and response categories, and these researchers effectively promoted generalization
skills generalized across nontraining scenarios, of social skills from training to nontraining
therapists, time, and settings. stimuli by (a) training with multiple exemplars
Similarly, Harris, Handleman, and Alessan- within a stimulus class (e.g., showing a variety
dri (1990) used prompting and reinforcement of reasons for sadness), and (b) training with
to teach 3 adolescent boys to display helping discriminative stimulus compounds (e.g., pre-
responses in a variety of situations. Although senting a combination of vocal and nonvocal
responding generalized from training to non- ‘‘sadness’’ stimuli).
training people and settings to some extent for In the present study, we sought to examine
each of the boys, levels of generalized respond- whether procedures used previously to teach
ing were lower than for training responding. other social skills could be extended to teaching
Expanding on the Harris et al. (1990) study, empathy skills to children with autism. Specif-
Reeve, Reeve, Townsend, and Poulson (2007) ically, the purpose of the present study was to
used video modeling, prompting, and rein- assess the extent to which a treatment package
forcement to teach 4 children with autism to consisting of the presentation of affective
emit helping responses in the presence of discriminative stimulus compounds, prompt
multiple discriminative stimuli from a variety delay, modeling through auditory scripts,
of helping categories (e.g., replacing broken manual prompts, behavioral rehearsals, and
materials, locating objects, putting away items). reinforcementwaseffectiveinteachingempathy
Thetreatment packagewas effectivein teaching skills in a pretend-play setting to children with
helping responses during training trials, and autism. Behavioral rehearsals and prompt delay
responding generalized across nontraining stim- (Charlop, Shreibman, & Thibodeau, 1985;
uli and response categories. Halle, Marshall, & Spradlin, 1979) were used
The aforementioned studies, like the present to prevent and correct errors, which were
study,approacheddeficitsinsocialbehaviorasa judged to be a potential difficulty. We also
stimulus control problem. Many behavioral assessed the extent to which empathy skills
deficits in individuals with autism have been generalized from training to nontraining probe
attributed to the phenomenon of stimulus stimuli and from training dolls and puppets to
overselectivity, the tendency to respond to a actual people in a nontraining setting.
TEACHING EMPATHY SKILLS 19
METHOD mostly of single words and short phrases, and
hissocialinitiationswereusuallynonvocal(e.g.,
Participants
smiling, walking toward a person). Like Jacob,
Participantswere4childrenwithautismwho
he did not demonstrate empathy responses to
attended an education program offering center-
displays of affect by others.
based and in-home behavioral intervention for
Ali was 8 years 9 months old at the start of
children with autism for 5.5 hr per day, 5 days
the study. She obtained a standard score of 40
perweek.Theywereidentifiedforparticipation
(age equivalent: 3 years 0 months) on the
based on anecdotal reports from their teachers
PPVT-III and 41 (age equivalent: 3 years 11
that they did not regularly demonstrate empa-
months) on the EVT. Her language consisted
thy toward others. All participants engaged in
mostly of single words and short phrases. She
minimal stereotypic or disruptive behavior and
rarely demonstrated empathy skills, although
had experience with token motivational sys-
she sometimes said ‘‘sorry’’ when she acciden-
tems. In addition, all participants had prereq-
tally bumped into another person.
uisite skills in vocal imitation of three-word
phrases modeled by an instructor and on
Setting and Materials
auditory recordings.
Sessions were conducted in a small room
Josh was 4 years 5 months old at the start of
furnished with small tables, chairs, and book-
the study. He obtained a standard score of
shelves. Toys were stored on the shelves, and
114 (age equivalent: 5 years 7 months) on the
dolls and puppets used for training and probe
Peabody Picture Vocabulary Test (3rd ed.,
trialswereplacedinseparatebinsinfrontofthe
PPVT-III) and 106 (age equivalent: 4 years 10
shelves. A Language Master was placed on one
months) on the Expressive Vocabulary Test
of the tables. A digital video camera was placed
(EVT). In addition to the behavioral interven-
in one corner of the room and was used to
tion program, he also attended an integrated
record some sessions.
preschool classroom for several hours per day,
spoke in full sentences with clear articulation, Vignettes and Target Responses
and established eye contact regularly during Affective discriminative stimulus compounds
socialinteraction.Althoughhesometimesmade containedonemotorandonevocalcomponent.
comments when observing displays of affect by They were presented in brief vignettes in which
others (e.g., ‘‘crying,’’ ‘‘sad’’), he rarely dis- an instructor held up a doll or puppet and
played appropriate empathy responses. pretended to make it say a statement paired
Jacobwas6yearsoldatthestartofthestudy. with an action. Vignettes fell into three
He obtained a standard score of 61 (age categories: sadness or pain, happiness or
equivalent: 2 years 9 months) on the PPVT- excitement, and frustration. For example, in
IIIand41(ageequivalent:2years4months)on the sadness or pain category, a vignette
the EVT. His language consisted mostly of consisted of the instructor holding a puppet,
requests, he rarely initiated social interaction, making it bump the table, and saying ‘‘ouch!’’
andhiseyecontactandattendingtootherswere as thoughthepuppet hadhurt itsleg.Vignettes
often poor. He demonstrated no empathy were assigned randomly to training and probe
responses to displays of affect by others. conditions. Sets of training and probe dolls and
Luke was 5 years 6 months old at the start of puppets were matched based on similarity (e.g.,
the study. He obtained a standard score of 56 each set contained boys, girls, and animals). To
(age equivalent: 1 year 10 months) on the promote generalization from training to non-
PPVT-III and 55 (age equivalent: 2 years 7 training stimuli, dolls and puppets were not
months) on the EVT. His language consisted assigned to specific vignettes.
20 JESSICA A. SCHRANDT et al.
Table 1
Training andProbeDiscriminative Stimuli andTraining Responses for the Sadness or Pain Category
Discriminativestimuli Responses
Training
‘‘Ouch!’’andbumpinglegontable ‘‘Areyouokay?’’andpattingarm
‘‘Idon’tfeelgood’’andsittingdown ‘‘Areyoualright?’’andpattingarm
‘‘Ihurtmyelbow’’andtouchingelbow ‘‘It’sokay’’andpattingarm
‘‘Mystomachhurts’’andrubbingstomach
‘‘Oww!’’andfallingdown
‘‘I’msoupset’’andsnifflingorcrying
‘‘Iburnedmyhand!’’andgrabbingtoastanddroppingit
Generalization
‘‘Ohno!’’andwipingeyeswithtissue
‘‘Icutmyfinger!’’andcuttingwithscissors
‘‘Ihaveaheadache’’andholdingforehead
Empathy was operationally defined as a doll or puppet. The prompter, seated behind
contextually appropriate response to a display the participant, waited for a designated interval
of affect by a doll, puppet, or person that foraresponseandthendeliveredaconsequence
contained motor and vocal components (in any (following training trials) or gave no feedback
order) and began within 3 s of the end of the (following baseline and probe trials). A baseline
display. For the previous example, the partici- orprobetrialendedwhen3 selapsed.Although
pant could be taught to say, ‘‘Are you okay?’’ 3 of the participants received training in only
and to pat the puppet’s arm. One motor the sadness or pain response category, baseline
response and three different vocal responses trialsfortheothertwocategorieswerepresented
were taught for each response category. To during each session. Dolls, puppets, toys, and a
promotegeneralitytonontrainingvignettes,the token motivational system were placed on the
responses were designed such that any response tablebetweentheparticipantandtheinstructor.
taught for a particular vignette was appropriate Priortoeachsession,participantschosefroman
for any vignette in the same response category. arrayofpreferredsnacksandactivitiesthatwere
In addition, other contextually appropriate delivered in exchange for every 10 tokens
responses not targeted during training were earned throughout the session (approximately
scoredascorrect.Forexample,inthesadnessor three times per session).
pain category, saying, ‘‘That’s too bad’’ while Baseline. Throughout the session, the
giving the doll a hug was scored correct even prompter delivered one token per trial for
though this response was not targeted directly. appropriate sitting and attending to the vi-
The same response was not, however, scored gnettes. Tokens were delivered randomly either
correct for the happiness or excitement catego- between trials as the instructor set up materials
ry, because it was not contextually appropriate. for the next vignette or directly before the
Acompletelistofvignettesandtargetresponses instructor presented the vignette. Instructors
for each response category is provided in and prompters did not deliver explicit instruc-
Tables 1, 2, and 3. tions or prompts to the participant regarding
the dolls, puppets, or toys on the table. Correct
Procedure responses were followed by a conversational
General procedure. Sessions were conducted exchange by the instructor (e.g., ‘‘thank you’’).
four to five times per week and lasted No feedback was given for incorrect responses.
approximately 20 to 30 min. A trial began Treatment package for training sessions. Train-
when the instructor presented a vignette with a ingsessionsconsistedof30trials(seventraining
TEACHING EMPATHY SKILLS 21
Table 2
Training andProbe Discriminative Stimuli andTraining Responses for the Happiness or ExcitementCategory
Discriminativestimuli Responses
Training
‘‘Thisisfun!’’andpushingcar ‘‘CanIsee?’’andmovinghandtowardownbody
‘‘Ifinished!’’andcoloringpicture ‘‘Showme!’’andmovinghandtowardownbody
‘‘Ididit!’’andcompletingpuzzle ‘‘Letmesee!’’andmovinghandtowardownbody
‘‘Watchme!’’andstretchingSlinky
‘‘Lookatthis!’’andshowinggiftbox
‘‘Iwon!’’andplayingHungryHungryHipposandputtingboth
handsintheair
‘‘Ifoundsomething!’’andshowinglizard
Generalization
‘‘Ilovethis!’’androllingAnimaniacsball
‘‘LookwhatIgot!’’andshowingMagicWheel
‘‘Thisisthebest!’’andplayingfrogpiano
and three nontraining probe-stimulus trials per prompter decreased the delay by 1 s if the
response category) presented in random order. participant responded correctly on fewer than
The prompter delivered manual and auditory fouroftheseventrainingtrials.Duringtraining
prompts according to a prompt-delay sequence trials, the instructor delivered tokens and
(Charlop et al., 1985; Halle et al., 1979) to behavior-specific praise following all correct
prevent errors. Specifically, the prompter used unprompted responses.
manual and auditory prompts immediately (0-s A behavioral rehearsal sequence was used for
delay) for three sessions when training was all training trials in the 0-s delay condition and
introduced for a particular response category. when the participant did not respond or
During the fourth session, the prompter used a responded incorrectly in all other prompt-delay
1-s delay. The prompter increased the delay conditions. First, the prompter manually
length by 1 s with a maximum of a 3-s delay prompted a correct motor response and simul-
following each session in which the participant taneously played an auditory script on a
respondedcorrectlyforatleastfouroftheseven Language Master, a device that plays aloud
trials in a particular response category. The phrases recorded on a strip of magnetic tape
Table 3
Training andProbe Discriminative StimuliandTraining Responses for the Frustration Category
Discriminativestimuli Responses
Training
‘‘Thiswon’twork!’’andputtingSylvestercarontable ‘‘Icanhelpyou’’andreachingonehandtowardpuppet/person,palm
turnedslightlyup
‘‘Ican’tdoit!’’andtryingtoputshapeinsorter ‘‘Wantsomehelp?’’andreachingonehandtowardpuppetorperson,
palmturnedslightlyup
‘‘Ican’tgetit’’andtryingtozipupcoat ‘‘I’llhelpyou’’andreachingonehandtowardpuppetorperson,palm
turnedslightlyup
‘‘Ohman!’’anddroppingbinofblocks
‘‘It’stoohard!’’andtryingtoputK’Nexpiecestogether
‘‘It’sstuck!’’andtryingtoremovelidfrombox
‘‘Ibrokeit!’’andholdingupbrokenLegomodel
Generalization
‘‘It’snotworking!’’andtryingtocolorwithmarkerwithcapon
‘‘Myarmisstuck!’’andtryingtoputonsweater
‘‘Thisistootight!’’andtryingtoopenjar
22 JESSICA A. SCHRANDT et al.
along the bottom of a card that is pushed Experimental Design
throughit.Iftheparticipantdidnotimitatethe A multiplebaselinedesign acrossparticipants
auditory script, the prompter played the card was used to assess the extent to which the
again and waited for a response. If the treatment package was effective in teaching
participant did not imitate the auditory script empathetic responding to displays of sadness or
after two presentations, the prompter provided pain to 4 participants (Josh, Jacob, Luke, and
vocal prompts (e.g., ‘‘Say ‘It’s okay’’’) until the Ali).For1participant(Josh),asecondmultiple
participant correctly imitated the script. Then, baseline design across response categories was
the prompter played the auditory script again used to assess the extent to which the treatment
and repeated this sequence until the participant package was effective in teaching empathetic
correctly imitated the auditory script immedi- responding across three response categories
ately after it was played on the Language (sadness or pain, happiness or excitement, and
Master. Once the participant imitated the frustration).
auditory script without prompts, the instructor
Data Collection and Interobserver Agreement
presented the same vignette, and the prompter
Theexperimenterandoneoffourinstructors
partially prompted the vocal and motor re-
conducted interobserver agreement assessments
sponses (e.g., played the first word of the
for at least 35% of sessions in each of the
auditory script and lightly tapped the partici-
experimental conditions for each participant.
pant’s arm). If the participant responded
Observers met a criterion of 80% agreement in
correctly, the prompter then gave the partici-
data collection before agreement measures
pant an opportunity to respond to the same
began. Independent observers sat at separate
vignettewithoutprompts. If the participantdid
tables and recorded motor and vocal responses
not respond or an error occurred, the sequence
verbatim. For each trial, they circled ‘‘yes’’ for
was repeated until the participant emitted a
responses that contained both a motor and
correct independent response.
vocal component, were contextually appropri-
Generalization to nontraining probe stimuli.
ate, and began within 3 s of the end of the
These trials were randomly interspersed
vignette, and ‘‘no’’ for responses that did not
throughout all training sessions (including
meet these requirements. Agreements were
baseline sessions); consisted of vignettes similar
defined as trials in which both observers circled
to those used during training trials; and
‘‘yes’’ or ‘‘no,’’ and disagreements were defined
incorporated dolls, puppets, and toys never
astrialsinwhichoneobservercircled‘‘yes’’and
associated with training. During these trials,
the other circled ‘‘no.’’ Agreements were
instructors and prompters followed the same
divided by agreements plus disagreements, and
procedures used during the baseline condition.
this ratio was converted to a percentage. Across
Generalization to nontraining people and
all participants and experimental conditions,
setting. These sessions occurred approximately interobserver agreement ranged from 83% to
once per week during baseline and training. 100%, indicating no more than one disagree-
Sessions consisted of seven trials per response ment per response category in a single session.
category. An actual person replaced the dolls
and puppets and presented the training vi- Procedural Integrity
gnettes used in the training setting. Sessions To ensure the integrity of the independent
were conducted in a large conference room not variable, procedural integrity measures were
associated with training. Again, instructors and collected during all interobserver agreement
prompters followed the same procedures used sessions. Independent observers assessed accu-
during the baseline condition. rate presentation of auditory and manual
TEACHING EMPATHY SKILLS 23
prompts as well as accurate delivery of tokens presented for the happiness or excitement and
for each trial during all conditions. Prompt frustrationresponsecategories.Becausevirtually
delivery was considered accurate if prompts no responding occurred during baseline, all
weredeliveredonlyafterthedesignatedamount baseline segments showed participants respond-
of time during the prompt-delay condition or ing incorrectly or not at all. All segments from
when participants responded incorrectly during treatment sessions showed responses that had
a training trial. Token delivery was considered been scored as correct. Observers were given a
accurate if it occurred only following attending written questionnaire and for each segment
behavior unrelated to the training task (e.g., circled‘‘yes’’or‘‘no’’forthequestion,‘‘Didthe
propersitting)duringbaselineorprobetrialsor child demonstrate empathy toward the doll or
following a correct unprompted response dur- puppet?’’
ing training trials.
For each trial, observers circled ‘‘yes’’ if
RESULTS
promptsandtokensweredeliveredaccuratelyor
‘‘no’’ if prompts and tokens were not delivered Figures 1 (Jacob, Luke, Ali) and 2 (Josh)
accurately. Agreements were defined as trials in depict the number of motor and vocal empathy
whichbothobserverscircled‘‘yes’’or‘‘no,’’and responses emitted to displays of affect with
disagreements were defined as trials in which training stimuli. During the baseline condition,
one observer circled ‘‘yes’’ and the other circled all participants responded infrequently to
‘‘no.’’ Agreements were divided by agreements displays of affect by the dolls and puppets.
plusdisagreements,andthisratiowasconverted Increased pretend-play empathetic responding
to a percentage. The mean percentage of occurredsystematicallywiththeintroductionof
accurate presentation of prompts and tokens treatment across participants.
was 99.9% across all participants and condi- Figures 3 (Jacob, Luke, Ali) and 4 (Josh)
tions (range, 98% to 100%). Mean interob- depict the number of motor and vocal empathy
server agreement on accurate presentation of responses emitted to displays of affect with
prompts and tokens was 99.9% (range, 98% to nontraining probe stimuli. Appropriate re-
100%). sponding generalized from training to non-
training vignettes, dolls, and puppets for all
Social Validity participants, increasing systematically with the
Nineteen teachers from the behavioral inter- introduction of treatment with training stimuli
vention program who were not involved in the across participants. For Josh and Jacob, gener-
study assessed the social validity of the results. alized motor and vocal responses emerged at
Observers watched videotaped segments of about the same time. Luke’s generalized motor
each of the participants responding to various and vocal responding with nontraining probe
vignettes with dolls and puppets. Each segment stimuli was variable, with vocal responding
(approximately 10 to 15 s in duration) showed emerging more quickly than motor responding.
the instructor presenting a vignette with a doll For Ali, the opposite was true; generalized
or puppet and the participant’s response, but motorrespondingemergedslightlymorequick-
did not depict instructor feedback. For all ly than generalized vocal responding.
participants, three baseline and three treatment Figures 5 (Jacob, Luke, Ali) and 6 (Josh)
segments were selected randomly from the last depict the number of motor and vocal empathy
three baseline and treatment sessions for the responses to displays of affect by Person 1 and
sadness or pain response category and were Person 2 in a nontraining setting. For Josh and
presented in random order. For Josh, three Jacob, motor and vocal responding generalized
baselineandthree treatmentsegmentswere also from training dolls and puppets to a teacher
24 JESSICA A. SCHRANDT et al.
Figure1. Numberofmotor(filledcircles)andvocal(opencircles)empathyresponsestodisplaysofsadnessorpain
withtraining stimulifor Jacob, Luke, andAli.
(Person 1) in a nontraining classroom. In Person1systematicallywiththeintroductionof
addition, both boys responded to displays of treatment with training stimuli in the training
affect by another teacher (Person 2) in the setting, although motor responding during this
nontraining setting, although no baseline mea- condition returned to baseline levels during the
surements were taken with this person (aside last two probe sessions. With the brief intro-
from a single observation in Session 25 in the duction of training (i.e., vignettes presented by
frustration category for Josh). Like Josh and Person 1, 3-s prompt delay, auditory prompts,
Jacob, Luke and Ali demonstrated generalized manual prompts, behavioral rehearsals, tokens
motor and vocal empathetic responding with delivered for correct responses) in the nontrain-
TEACHING EMPATHY SKILLS 25
Figure 2. Number of motor (filled circles) and vocal (open circles) empathy responses to displays of affect with
training stimulifor Josh.
ing setting (Sessions 55 and 56), motor to Person 2 were also observed for both Luke
responding increased to the same high level as and Ali, although, as with Person 1, vocal
vocal responding for both Luke and Ali. responding was emitted at higher levels than
Increases in generalized empathetic responding motor responding.
26 JESSICA A. SCHRANDT et al.
Figure3. Numberofmotor(filledcircles)andvocal(opencircles)empathyresponsestodisplaysofsadnessorpain
withnontraining probestimuli for Jacob, Luke,andAli.