JASH copyright 1998 by

1998, Vol. 23, No.4, 300-308 The Association for Persons with Severe Handicaps

Increasing the Social Behavior of Young

Children With Autism Using Their
Obsessive Behaviors
Mary J. Baker, Robert L. Koegel, and Lynn Kern Koegel
University of California, Santa Barbara

This study systematically asked whether individual The early intervention studies on social interactions
topics or themes on which children with autism perse- of children with autism focused on teaching specific
verated across a variety of settings (often called obses- social skills (Beck & Forehand, 1984; Pellegrini & Ur-
sions) could be used to create the theme for a socially bain, 1985; Williams, 1989). Many of these studies re-
appropriate game. Data collected within the context of a lied on creating special social skills training groups and
multiple baseline design revealed very low levels of so- teaching individual skills such as appropriate eye con-
cial interaction during play periods in the baseline con- tact, turn taking, sharing, and asking for help (Kamps et
dition. In contrast, when the children with autism were aI., 1992; Koegel & Frea, 1993; Taras, Matson, & Leary,
taught a socially appropriate game (e.g., one child who 1988). Other strategies, which resulted in increases in
perseverated on maps was taught a tag game played on a social interactions and play in children with autism, fo-
giant outline of a us map), the percent of social inter- cused on teaching typically developing peers to initiate
actions increased dramatically and continued to be high play with children with autism (Brady, Shores, McEvory,
during follow-up measures. Generalization measures in- Ellis, & Fox, 1987; Ragland, Kerr, & Strain, 1978;
dicated that following intervention, the children also Strain, Kerr, & Ragland, 1979). Although such methods
demonstrated increases in social interaction during other revealed increases in the interaction of children with
play activities. The results are discussed in terms of in- autism and their peers, generalization to nontraining
corporating obsession themes into common games to situations and peers continued to be problematic
create intrinsically reinforcing appropriate activities for (Odom & Strain, 1984).
increasing social interactions, and in relation to develop- The need for social reciprocity for the development
ing activities that capitalize on the child with autism's of pragmatically appropriate social interactions led re-
interests, so that the child is a valued member of the peer searchers to examine the effects of teaching children
group. with autism to initiate interactions with their nondis-
abled peers (Haring & Lovinger, 1989). Furthermore,
DESCRIPTORS: autism, obsessions, generalization Oke and Schreibman (1990) found that social interac-
measures, social behavior tions not only increased, but disruptive behaviors also
decreased with this type of intervention. Gaylord-Ross,
One of the defining criteria for autism is extreme Haring, Breen, and Pitts-Conway (1984) conducted a
difficulty with social behavior. Even when social skills study to increase social interactions by teaching autistic
improve while some individuals interact with adults, adolescents to initiate social interactions by sharing
many young children with autism continue to exhibit highly desired age-appropriate objects with peers. The
little or no social interactions with peers (Scott & Gil- objects were used to promote social encounters and
liam, 1987). They may actively avoid social contact with were selected to be reinforcing to both the autistic stu-
others, adhere to rigid schedules, and inappropriately dents and the nondisabled peers. The results showed
perseverate on objects or topics. increases in the frequency of initiations, the duration of
interactions, and generalization across peers (but not
This research was supported in part by PHS grant MH28210 settings). The use of preferred objects may have helped
from the National Institute of Mental Health, and by U.S.
increase the intrinsic motivation to socially interact for
Department of Education grant G0087C0234.
The authors acknowledge Holly Reynolds, Heather Han- both the adolescents with autism and typically devel-
sen, and Julie Hernandez, who assisted with the data collection oping adolescents. Another study that employed child
and videotape analyses. The authors also are grateful to Dr. preference of conversational topics and play activities
Mike Furlong for his assistance. to increase autistic children's social behavior was con-
Address all correspondence and requests for reprints to
Robert L. Koegel, Ph.D., Counseling/Clinical/School Psychol- ducted by Koegel, Dyer, and Bell (1987). They found a
ogy Program, Graduate School of Education, University of negative correlation between child-preferred activities
California, Santa Barbara, CA 93106-9490. and social avoidance and demonstrated that children
300

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

 Social Play and Obsessions in Autism 301

can be taught to initiate child-preferred activities in did not cause increases in the perseverative behavior
community settings, thereby reducing social avoidance during other times of the day.
behaviors. This study extends this line of research by analyzing
Such studies demonstrate that, under controlled con- whether obsessive behaviors could be used functionally
ditions within the context of a community setting, pro- as the theme of an appropriate social game, and wheth-
cedures can be implemented for promoting interactions er the game in turn would increase social interactions
between children with autism and their typically devel- between children with autism and their nondisabled
oping peers (Odom, Hoyson, Jamieson, & Strain, peers. The literature suggests that the use of an autistic
1985). Systematically building on this research has been child's perseverations incorporated into social play in-
the addition of a greater number of naturalistic tech- teractions could be successful in two ways. First, it may
niques (those that are conducted in more loosely con- produce interest in the appropriate social activity and
trolled contexts) to improve the generalization and du- create positive affect. Second, because of the intrinsic
rability of social interactive play with peers (Pierce & reinforcement the obsessive behavior may introduce
Schreibman, 1995). into the appropriate social game, generalization and
Specifically included in such studies were motiva- maintenance may be more likely to occur. Specifically,
tional components, such as child choice of stimulus ma- we asked whether an intervention based on incorporat-
terials, interspersing of maintenance trials (tasks the ing an autistic child's obsessive behavior into an appro-
child has already mastered) among new acquisition tri- priate social game would (1) increase the percent of
als, using natural reinforcers that are directly related to time engaged in social interactive play; (2) result in
the child's response, and reinforcing attempts at social generalization and maintenance to other social interac-
interactions and functional play (cf. Koegel, O'Dell, & tive play; and (3) result in increased positive affect dur-
Koegel, 1987). These components may have been par- ing the social interactions for both children with autism
ticularly important because, in general, children with and for their typically developing peers.
autism seem very unmotivated (Koegel & Egel, 1979).
Therefore, incorporating these components into the Method
teaching situations appears to motivate the children to Participants
socially interact and initiations often occur spontane- Three children who were diagnosed with autism, ac-
ously. cording to criteria outlined in the 4th ed. of the Diag-
In contrast to their lack of motivation to engage in nostic and Statistical Manual of Mental Disorders
social interactions, one area where many children with (DSM-IV; American Psychiatric Association, 1994), by
autism seem to be especially motivated relates to idio- two independent agencies participated in this study. In-
syncratic objects, topics, and themes on which a child dividual child descriptions are as follows.
perseverates across multiple settings (often called ob- Don was 7 years and 7 months old at the start of the
sessions). Some have hypothesized that these obses- study. He participated in a full-inclusion second-grade
sions may be very reinforcing forms of self-stimulatory class in a public elementary school with assistance from
behavior which oftentimes are problematic and dys- special education support personnel. He participated in
functional (Epstein, Taubman, & Lovaas, 1985). These most of the academic activities during classtime. How-
obsessions seem to be child preferred and generally ever, he frequently engaged in off-task behavior, pri-
occur at a high frequency (Lovaas, Newsom, & Hick- marily consisting of incessant repetitive speech about
man, 1987). Although typically viewed as abnormal, certain obsessive topics (especially U.S. states). He gen-
some researchers have shown that such behaviors can erally played in isolation at recess and lunch periods,
be incorporated effectively into intervention programs with his activities primarily consisting of sitting on top
that increase appropriate behaviors without increasing of the jungle gym apparatus and displaying stereotypic
the time engaged in the obsessive behavior (Wolery, behavior (blank stare and finger flapping). His Vine-
Kirk, & Gast, 1985). For example, perseverative or ste- land Adaptive Behavior Scales adaptive behavior com-
reotyped ritualistic behaviors have been used as rein- posite score was 65. His overall IQ score on the Wech-
forcers to improve correct responding (Wolery, 1978; sler Intelligence Scale for Children (WISC-III) was 83
Wolery et aI., 1985), to increase the use of spontaneous (verbal = 65, performance = 106). He was able to
appropriate sentences (Hung, 1978), to increase prevo- combine four to five words to form grammatically cor-
cational work rates (Sugai & White, 1986), to increase rect sentences, but he rarely engaged in social commu-
correct task performance (Charlop, Kurtz, & Casey, nication.
1990), and to decrease off-task, stereotypy, aggression, Amy was 5 years and 4 months old at the start of the
and tantrum behaviors (Charlop-Christy & Haymes, study. She attended a full-inclusion preschool during
1996). This approach allows behaviors that are difficult the initial portion of the baseline condition. She com-
to eliminate to be used as an advantage to teach new pleted the baseline in a full-inclusion public school kin-
adaptive responses. Further, Charlop et al. (1990) also dergarten class with the assistance of special education
noted that using the obsessive behaviors as a reinforcer support personnel. She engaged in most of the same

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

302 Baker, Koegel, and Koegel

academic activities as her peers, but she did not socially out the inclusion of the obsession theme (e.g., "Don,
interact with the other children. She generally exhibited let's play tag," or "Amy, let's play follow the leader").
disruptive behavior, such as yelling or loudly saying Throughout all phases of the study, none of the partici-
phrases such as "Go away!" when other children ap- pants were aware of the hypotheses or the experimental
proached her. Her free-time periods primarily con- conditions.
sisted of playing in isolation with dolls and Disney char- Intervention. To determine each child's obsession
acters, swinging, or wandering around. In addition, theme to be incorporated into the playground games,
Amy persistently engaged in various repetitive behav- an obsession topic was chosen following interviews with
iors relating to Disney characters, such as insisting on each child's teacher, teacher's aide, and parent, and
wearing Disney clothing and repeatedly asking for Dis- informal child observations. For the obsession topic to
ney paraphernalia. She obtained a Vineland Adaptive be chosen, the topic had to be considered by all of the
Behavior Scales adaptive behavior composite score of observers and interviewees to be abnormally persevera-
65. Amy could speak in simple three to four-word sen- tive. Each obsession theme was then incorporated into
tences but rarely initiated conversations and usually did a common playground game in accordance with the
not respond to others' initiations. rules of the playground, the age of the children, and
Jenny was 8 years and 9 months old at the start of the without limitation to the number of children who could
study. She was enrolled in a special education fourth- participate.
grade class with part-time mainstreaming. Her aca- Intervention consisted of an adult prompting each
demic performance was higher than her peers in her target child, as well as any group of nondisabled peers
special education class, but about 1 year below her age who showed interest and approached the activity, to
level in the regular education classroom. She tended to play a game that was commonly played at school, but
play in isolation at recess periods, typically swinging on which was modified to include the autistic child's ob-
a particular swing. She cried and threw tantrums when sessive behavior as a theme. Participation in the game
other children requested that she share the swing. At was voluntary (participants varied from day to day) and
home and school Jenny demonstrated obsessions with no extrinsic reinforcers were provided. There usually
movies. She had seen and memorized virtually every were enough children to play the game. The nondis-
current child-oriented movie, knew all the characters' abled children often initiated the game. On occasions
names, and continuously repeated exact scripts of the when participants were not readily available in the
movies. Her reported IQ on the Leiter International game area, the child with autism was prompted to re-
Performance Scale was 107. Her age-equivalent scores cruit peers to engage in the game (e.g., to say, "Let's
on the Peabody Picture Vocabulary-Revised Test and play the [obsession theme] Game"). The above types of
the Expressive One Word Picture Vocabulary Test prompting were performed one to two times during the
were 6 years 3 months and 6 years 10 months, respec- initial days of intervention, fading during the remaining
tively. Jenny engaged in lengthy monologues related to 1 to 2 days of the intervention phase. At that point, and
movies with adults, but rarely engaged in appropriate for the remainder of the study, all of the children's
social interactions with peers. selection of and participation in games and activities
was entirely independent and voluntary, with no adult
Design and Procedures prompting.
A multiple baseline across participants design (Bar- Child 1. Don's obsession theme was the geography of
low & Hersen, 1984) was employed to assess whether the United States (e.g., he had memorized the location
an intervention based on incorporating an autistic of every state and capital in the country). This obsessive
child's obsession theme into an appropriate social game topic was incorporated into a tag game that took place
would increase the time and quality of social interactive on a giant outline drawing of a map of the United States
play. Data were collected in randomly selected 10- located on the outside playground. The game consisted
minute probes, twice weekly, during regular recess or of one "caller" (either the child with autism or a non-
lunch periods at each child's school setting during the disabled peer) who would yell out the name of a state
baseline, intervention, and follow-up conditions. and indicate how the other children should travel there
Baseline. During the baseline condition, the children (i.e., hopping, running, walking, twirling). After the
participated in activities during their regular recess and "call" (e.g., "hop on one foot to Maine" or "run to
lunch settings without any special manipulations or in- California"), the other children would try to travel to
structions. The children were free to engage in any the state without getting tagged by the caller. If a child
games or activities. As a control for the initial prompt- was tagged then she became the new caller. If no one
ing to start the games during the intervention sessions, was tagged, the caller called out another state and the
two or three baseline sessions were selected for each traveling mode and continued to do so until someone
child (see arrows in Fig. 1). Adults and peers were was tagged.
asked to prompt the child with autism to play the same Child 2. Amy's obsession theme was obtaining and
games that would be used in the intervention, but with- repeatedly interacting with paraphernalia relating to

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

 Social Play and Obsessions in Autism 303

Disney characters. The Disney theme was incorporated interactions were defined as playing alone; engaging in
into a follow the leader game. The game was intro- inhibiting stereotypic behavior; engaging in ritualistic
duced to the children as the "Follow Mickey Game." It behaviors related to the obsession theme; and/or not
consisted of each child picking out and wearing a Dis- responding appropriately or not attending to the other
ney character hat or laminated emblem (i.e., Mickey children during a game.
Mouse, Minnie Mouse, Goofey, or some other charac- Time engaged in social play interactions was calcu-
ter hat or emblem). The child wearing the Mickey lated as the percent of 10-sec intervals with social play
Mouse emblem was the leader. That child could go on interactions during each 10-minute probe. A full inter-
any play equipment in the kindergarten play yard or val scoring system was utilized. Additionally, an ob-
travel around the play area in any manner (i.e. run, hop, server unfamiliar with the experimental hypotheses or
or skip) and all the other children were to follow and conditions rated the target child and a composite of the
imitate the leader. After a circuit of the playground, peers (based on the average behavior of two different
any participating child could call out "change charac- randomly selected peers at each session) on two 6-point
ters." Upon hearing that request, each child was re- scales relating to interest and happiness (based on the
quired to trade emblems with someone else. Again, the Dunlap and Koegel Affect Scale, 1980) at the end of
new child with the Mickey Mouse emblem would lead each 10-minute probe. The scales ranged from zero
the group. (disinterested/ unhappy) to five (interested/happy).
Child 3. Jenny's obsession theme was movies (i.e., Table 1 shows the specific scales.
she memorized every current movie title, characters, Reliability. Two observers, at least one of whom who
and much of the scripts). Movie themes were incorpo- was naive to the purpose of the study, independently
rated into a tag game, using 10 to 15 laminated photo- recorded each child's percent of intervals with appro-
copies of pictures on current child-oriented rental priate social play interactions in vivo or from video-
movie boxes. The movie pictures were placed around a tapes. Reliability was computed on an interval-by-
large outside grassy play area and were considered to interval basis for approximately 33 % of the sessions in
be "bases" of safety. The object of the game was for a all conditions for each child. An agreement was
caller to quote something from one of the movies, name counted when both of the observers recorded an inter-
a character, or describe a scene. Then the other chil- val with the target child engaging in appropriate social
dren needed to run to that movie to be safe from being play interactions or when both observers recorded an
tagged by the caller. If a child was tagged, that child interval with the target child not engaging in appropri-
became the caller. If no one was tagged, the caller ate social play interactions. Disagreements were de-
would continue. fined as one of the observers recording the target child
Maintenance. The maintenance condition was a re- engaging in appropriate social play during an interval
turn to the baseline condition (without any adult initia- and the other observer not recording the target child
tions). This occurred in the same outside play setting, engaging in appropriate social play during the interval.
with the exception of the availability of the new game Reliability was calculated for each session using the
materials where relevant (e.g., placards). As in the formula: number of agreements divided by the number
baseline, the children were free to engage in any games of agreements plus disagreements times 100. Reliability
or activities. was also computed for 200/0 of the sessions for the affect
Follow-ups. Follow-up sessions were conducted ap- measure. Agreements for recording affect were defined
proximately 1 and 2 months after the end date of the as both observers recording within the same category
maintenance condition to assess long-term mainte- (positive, neutral, or negative).
nance of appropriate social play interactions. All con- The average percentage agreement for appropriate
ditions were exactly the same as in the baseline and social play interactions for Don was 94 % with a range
maintenance condition, with no adult initiations. from 80% to 100%. The average percentage agreement
for Amy for appropriate social play interactions was
Dependent Measures 97% with a range of 870/0-1000/0. The average percent-
Data were collected for two dependent measures age agreement for Jenny was 94°k with a range of78%-
throughout the study: (1) time engaged in social play 1000/0. Percentage agreement for the affect measure for
interactions in obsession and nonobsession theme Don was 1000/0, for Don's peer composite it was 75%,
games and (2) subjective ratings of child affect. Appro- for Amy it was 1000/0, for Amy's peer composite it was
priate social play interactions were defined as the child 83%, for Jenny it was 750/0, and for Jenny's peer com-
with autism actively participating in a game or social posite it was 75%.
activity; being attentive to the game or social activity;
playing with the other children, and not just with an Results
object; exhibiting reciprocal turn taking as per the rules
of the game or activity; and not ritualistically engaging Figure 1 shows the percentage of intervals with the
in obsessive behaviors. In contrast, inappropriate social target children engaged in appropriate social play in-

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

304 Baker, Koegel, and Koegel

Table 1
Rating Scale for Child Affect
INTEREST
Disin terested Neutral Interest Interested
Child looks bored, noninvolved not Neither particularly interested nor Attends readily to task: responds readily
curious or eager to continue activity. disinterested. Child seems to passively and willingly. Child is alert and
May yawn or try to avoid the accept situation. Doesn't rebel but is involved in activity. (Score 4 or 5,
situation. Spends much time looking not eager to continue. (Score 2 or 3, depending on level of alertness and
around and not attending to task. If depending on extent of interest) involvement)
child does respond may be long
response latency. (Score 0 or 1,
depending on extent of disinterest)
I I -------- I --------- I --------- I --------1
o 1 2 3 4 5

HAPPINESS
Unhappy Neutral Happy
Doesn't appear to be decidedly happy
Cries, pouts, tantrums, appears to be Smiles, laughs approximately seems to
or particularly unhappy. May smile or
sad, angry, or frustrated. Child seems be enjoying self. (Score 4 or 5,
not to be enjoying self. (Score 0 or 1
frown occasionally but overall, seems depending on extent of enjoyment)
rather neutral in this situation. (Score
depending on extent of unhappiness)
2 or 3, depending on extent of
happiness)
I -------- I --------1 I --------- I I
o 1 2 3 4 5

teractions during recess/lunch periods. In the baseline obsession theme games with appropriate social interac-
condition, all 3 children exhibited low levels of social tions occurring during an average of 800/0 (range =
play interactions. For example, Don socially interacted 690/0-870/0) of the intervals (indicated by open circles).
with his peers an average of 19% (range = 0°k>-55 % ) Amy also continued to play her obsession theme on
of the intervals for 7 sessions. Amy rarely socially in- some days and averaged 70% (range = 480/0- 880/0) of
teracted with her peers, averaging only 20/0 (range = the intervals with appropriate social interactions (indi-
0°k>-10%) of the intervals with social play for 10 ses- cated by closed circles). Similar to Don, on days when
sions. Similarly, Jenny averaged 150/0 (range = 00/0- Amy did not engage in the obsession theme game, she
370/0) of the intervals with social play interaction with engaged in nonobsession theme games with appropri-
her peers for 13 sessions. ate social interactions occurring during an average of
The data show that during the intervention, mainte- 560/0 (range = 500/0-620/0) of the intervals (indicated by
nance, and follow-up phases, all 3 children showed in- open circles). Last, Jenny also continued to play the
creases in their appropriate social behavior. Specifi- obsession theme on some days, averaging 770/0 (range
cally, during intervention, Don averaged 660/0 (range = = 500/0 -95 ok> ) of the intervals with appropriate social
60 % - 7 5 % ) of the intervals with appropriate social play interactions (indicated by closed circles). In addition,
interactions whereas Amy averaged 760/0 (range = on days when she did not engage in the obsession
60°k>-87 %). Jenny immediately began engaging in ap- theme game, she engaged in nonobsession theme
propriate social play interactions during 1000/0 of the games with appropriate social interactions averaging
intervals and continued at this level throughout the in- 970/0 (range = 930/0-1000/0) of the intervals (indicated
tervention phase. by open circles).

Maintenance Follow-up
To assess whether the children's appropriate social The children continued to engage in appropriate so-
behavior would maintain in the absence of the adult cial play interactions during 1- and 2-month follow-up
who introduced the game, a maintenance condition was sessions. Specifically, Don averaged 690/0 (range =
conducted without the presence of the adult who taught 510/0-83 0/0) of the intervals with appropriate social play
the game. Under these conditions, Don continued to interactions during the follow-up sessions. Further,
play the obsession theme game on some days, averag- Don's appropriate social play interactions occurred on
ing 880/0 (range = 70%-100%) of the intervals with other days, averaging 71°k> (range = 51%-830/0) of the
appropriate social interactions (indicated by closed intervals (indicated by open circles) during nonobses-
circles). Further, on other days when Don did not en- sion theme games. Amy averaged 800/0 (range = 530/0-
gage in the obsession theme game, he played other non- 950/0) of the intervals with appropriate social play in-

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

 Social Play and Obsessions in Autism 305

Baseline Intervention Maintenance Follow Up Follow Up

I

100 : ,..../',
. .. ObsessionTheme
90
:j . \ Game

V! ~on-ObseSSion I
80

'o"
C 70
60
:;;
..
u
CO
50 \

.=...
40 : Theme Game
CD
I
30 I
I
20 I
I
10 I
I
o I
I

ca
'(3
o 100
in

\
90
CD 80
C; 70
'i:

..
0.
o
0.
0.
60
50
40
<C 30

...
.c 20 preschool kindergarten

'i 1~~Jo
ca'" t
.. t+
>

.=...
CD

....o
100
90
...
...c 80

»v
70

..
CD
u
CD
Q.
60
50
40

20
30
10
o
t--r--r-.,.........,-.--r-r-r-""-+-"""'r""......r-""'"T"""'I"""T"""I"""T""""I"""'T'"'"I"""T".......,.--T.........,--,r-r-r-o-,............--~...............

12345678 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

tt t
Initiation Probe
Weeks
Figure 1. The percentage of intervals the children with autism engaged in appropriate social play interactions during recess and lunch
periods during the baseline, intervention, maintenance, and follow-up conditions. Closed circles, obsession theme games; open
circles, nonobsession theme games.

teractions. Amy also engaged in appropriate social in- tion of the two dimensions, an average was obtained for
teractions during nonobsession theme games, exhibiting each 10-minute session to form composite ratings of
appropriate social play interactions in 82°ib of the in- affect. In Figure 2, a rating of 3.3 to 5 indicates a posi-
tervals during the 2-month follow-up session (indicated tive score (i.e., very interested, very happy), 1.71 to 3.29
by open circle). Finally, Jenny averaged 840/0 (range = indicates neutral affect (i.e., neither interested nor dis-
77%-870/0) of the intervals with appropriate social play interested/neither decidedly happy nor particularly un-
interactions during the follow-up sessions. Similar to happy), and 0 to 1.7 indicates a negative score (i.e.,
Don, Jenny's appropriate play interactions occurred on disinterested/unhappy). The data show that during the
other days during nonobsession games, averaging 86 % intervention condition, their ratings of affect increased
(range = 85 % - 8 7 % ) of the intervals (indicated by as did the affect of their peers. These increases in affect
open circles). were also evident during the maintenance and follow-
up sessions.
Don's average affect score during the baseline con-
Ratings of Affect dition was 2.9 (neutral) and the average peer composite
The target children and a peer composite were was 3.2 (neutral). During the intervention condition,
scored on two dimensions of affect (interest and hap- Don's average affect score increased to 4.5 (positive)
piness) related to the games. Due to the high correla- and the average peer composite affect score was 4.4

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

306 Baker, Koegel, and Koegel

Baseline Intervention Maintenance Follow Up Follow Up

'-------',

Q)

o
.~ 4:
~ 3.5
N !
:
7'l'1J. \
o
tJ)

2.5

1.5

0.5

.,

Weeks
Figure 2. The average affect scores for the children with autism (closed circles) and their peers (open circles) during baseline,
intervention, maintenance, and follow-up conditions.

(positive). The positive affect continued for both Don tive) during the maintenance condition and 4.8 (posi-
and his peers during the maintenance and follow up tive) at follow-up.
conditions. Don's average affect score during the main- Similar to Don, Jenny showed an increase in average
tenance sessions was 4.5 (positive) and it was 3.6 (posi- affect scores from neutral during the baseline condition
tive) during the follow-up sessions. The peer composite to positive during intervention, maintenance, and fol-
average affect score during the maintenance condition low-up. Jenny's average affect score during baseline
was 4.6 (positive) and during the follow-up condition it was 2.0 (neutral) and the peer composite was 2.5 (neu-
was 4.25 (positive). tral). Jenny's average affect score increased to 4.8 (posi-
Amy and her peers' affect also increased from neu- tive) and the peer composite to 4.2 (positive) during the
tral during the baseline condition to positive during in- intervention condition. The positive affect continued
tervention, maintenance, and follow-up. Amy's base- during the maintenance and follow-up, averaging 4.65
line average affect score was 1.7 (negative) and the peer (positive) and 4.7 (positive), respectively. Likewise, her
composite was 2.75 (neutral). During the intervention peer composite average affect score was 4.8 (positive)
condition, Amy's average affect score increased to 3.5 during the maintenance and 4.8 (positive) at follow-up.
(positive) and the peer composite was 4.75 (positive).
Amy's average affect was 4.4 (positive) during the Discussion
maintenance condition and 4.25 (positive) at follow-up. Results of this investigation showed that the obses-
Her peer composite average affect score was 4.8 (posi- sion themes of children with autism, which are typically

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

 Social Play and Obsessions in Autism 307

viewed as problematic, can be transformed successfully into the game. For example, after the intervention, the
into common games to increase positive social play in- peers commented on the positive skills that the children
teractions between children with autism and their typi- with autism exhibited (e.g., "He's the smartest kid in
cally developing peers in inclusive school environ- the school, he knows all of the states"). This perceived
ments. These increases in appropriate social play inter- value as a contributing member of the social group may
actions maintained in the absence of the adult who have been a factor in the high level of social play inter-
initially prompted the game. Further, the appropriate actions following intervention and during nonobsessive
social play generalized to other nonobsession theme theme games. This is consistent with Guralnick's (1990)
games following intervention. Additionally, both the suggestion that the competence of children during so-
children with autism and their nondisabled peers' affect cial interactions, and not the fact that they have a dis-
increased positively, showing improved interest and ability, is the key variable to improving social relation-
happiness following intervention. ships and their status among peers. Future research to
These results are interesting in light of several issues. systematically address these variables may enhance our
First, although the children with autism and their peers knowledge of these issues.
were prompted to engage in playground games during Also, it is interesting to note that anecdotal com-
baseline, this did not seem to be sufficient to result in ments by the parents and teachers of the children par-
improvements in their social behavior. However, fol- ticipating in the study suggested no increases in the
lowing incorporation of the obsession theme into the autistic children's obsessive behaviors as an undesirable
games, the children with autism appeared to be highly side effect of the intervention. In fact, for two of the
motivated to engage in and maintain appropriate social children, the parents and teachers reported what ap-
interactions. These findings are consistent with re- peared to be a decrease in obsessive behavior following
search showing that as children with autism improve the initiation of the games. This is consistent with Char-
with intervention programs, their stereotypy or self- lop et al. (1990), who suggested that when obsession
stimulatory behaviors change from low-level typologies themes were used as a reinforcer for appropriate be-
(e.g., rocking, spinning) to higher-level typologies (e.g., havior, they did not increase during other times of the
verbal perseverations, conversational obsessions), yet day. The present study showed increases in appropriate
maintain their very strong reinforcing properties (Char- social interactions (which by definition did not include
lop et al., 1990; Epstein et al., 1985; Lovaas et al., 1987). obsessive ritualistic behaviors). However, it may be in-
The result showing improvements in affect for both teresting for future research to measure obsessive ritu-
the children with autism and their nondisabled peers is alistic behaviors directly in order to assess possible in-
interesting. Several researchers have suggested that to verse relationships.
develop relationships between children with and with- The post comments by the parents and teachers sug-
out disabilities, activities need to be mutually reinforc- gest that appropriate behaviors actually may have re-
ing (Hurley-Geffner, 1995; Rogers & Lewis, 1989). The placed some of the children's inappropriate obsessive
improvements in affect in the present study suggest that behaviors. That is, the children were reported to inter-
the games were mutually reinforcing for all the children act socially more with siblings, analogous to the ob-
involved. More controlled social validation data may be served increase in appropriate social behavior with
important for future research in order to assess the mu- peers following the intervention. It may be useful for
tual reinforcement properties of the games. Second, future research to assess the generalization effects
considering social reciprocity may have been important across varied settings such as in the classroom and
so that both the children with autism and their nondis- home environments. Additional research in this area
abled peers had equitable roles in the relationship may prove profitable for understanding the nature and
rather than "helper/helpee" roles (Hurley-Geffner, functions of obsessive and play behaviors and for in-
1995; Strain, Odom, & McConnell, 1984; Strain & creasing positive interactions among children with and
Shores, 1977; Van der Klift & Kunc, 1994). In the pre- without disabilities.
sent study, the games were designed to be interesting
and age appropriate and to incorporate strengths that References
the children with autism could bring to the game. Thus,
American Psychiatric Association. (1994). Diagnostic and sta-
the children with autism appeared to be inherently val-
tistical manual of mental disorders (4th ed.) Washington,
ued members of the games and did not need additional DC: Author.
assistance from their peers. The positive affect results Barlow, D. H., & Hersen, M. (1984). Single case experimental
and anecdotal comments made by the peers support the designs: Strategies for studying behavior change (2nd ed.)
notion that reciprocal meaningful relationships were New York: Pergamon Press.
formed between the children with autism and their non- Beck, S., & Forehand, R. (1984). Social skills training for chil-
dren: A methodological and clinical review of behaviour
disabled peers. Also, the children with autism in this modification studies. Behavioural Psychotherapy, 12, 17-45.
study were viewed as socially competent due to their Brady, M. P., Shores, R. E., McEvory, M. A., Ellis, D., & Fox,
expertise related to the obsession theme incorporated J. J. (1987). Increasing social interactions of severely handi-

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016

308 Baker, Koegel, and Koegel

capped autistic children. Journal of Autism and Develop- Oke, N. J., & Schreibman, L. (1990). Training social initiations
mental Disorders, 17, 375-390. to a high-functioning autistic child: Assessment of collateral
Charlop, M. H., Kurtz, P. F., & Casey, F. G. (1990). Using behavior change and generalization in a case study. Journal
aberrant behaviors as reinforcers for autistic children. Jour- of Autism and Developmental Disorders, 20, 479-497.
nal of Applied Behavior Analysis, 23, 163-181. Pellegrini, D. S., & Urbain, E. S. (1985). An evaluation of cog-
Charlop-Christy, M. H., & Haymes, L. K. (1996). Using obses- nitive problem solving training with children. Journal of
sion as reinforcers with and without mild reductive proce- Childhood Psychology and Psychiatry, 26, 17-41.
dures to decrease inappropriate behaviors of children with Pierce, K., & Schreibman, L. (1995). Increasing complex social
autism. Journal ofAutism and Developmental Disorders, 26, behaviors in children with autism: Effects of peer-
527-546. implemented pivotal response training. Journal of Applied
Dunlap, G., & Koegel, R. L. (1980). Motivating autistic chil- Behavior Analysis, 28, 285-295.
dren through stimulus variation. Journal of Applied Behav-
Ragland, E. U., Kerr, M. M., & Strain, P. S. (1978). Behavior
ior Analysis, 13, 619-627.
of withdrawn autistic children: Effects of peer social initia-
Epstein, L. J., Taubman, M. T., & Lovaas, O. I. (1985).
tions. Behavior Modification, 2, 565-578.
Changes in self-stimulatory behaviors with treatment. Jour-
nal of Abnormal Child Psychology, 13, 281-294. Rogers, S. J., & Lewis, H. (1989). An effective day treatment
Gaylord-Ross, R. J., Haring, T. G., Breen, C., & Pitts-Conway, model for young children with pervasive developmental dis-
V. (1984). The training and generalization of social interac- orders. Journal of the American Academy of Child and Ado-
tion skills with autistic youth. Journal of Applied Behavior lescent Psychiatry, 28, 207-214.
Analysis, 17, 229-247. Scott, B., & Gilliam, J. (1987). Curriculum as a behavior man-
Guralnick, M. (1990). Social competence and early interven- agement tool for students with autism. Focus on Autistic
tion. Journal of Early Intervention, 14, 3-14. Behavior, 2, 1-8.
Haring, T. G., & Lovinger, L. (1989). Promoting social inter- Strain, P. S., Kerr, M. M., & Ragland, E. U. (1979). Effects of
action through teaching generalized play initiation re- peer mediated social initiations and prompting/
sponses to preschool children with autism. Journal of The reinforcement procedures on the social behavior of autistic
Association for Persons with Severe Handicaps 14, 58-67. children. Journal ofAutism and Developmental Disorders, 9,
Hung, D. W. (1978). Using self-stimulation as reinforcement 41-54.
for autistic children. Journal of Autism and Childhood Strain, P. S., Odom, S. L., & McConnell, S. (1984). Promoting
Schizophrenia, 8, 355-367. social reciprocity of exceptional children: Identification, tar-
Hurley-Geffner, C. M. (1995). Friendships between children get behavior selection, and intervention. RASE: Remedial
with and without developmental disabilities. In R. L. Koegel and Special Education, 5, 21-28.
& L. K. Koegel (Eds.), Teaching children with autism: Strat- Strain, P. S., & Shores, R. E. (1977). Social reciprocity: A re-
egies for initiating positive interactions and improving learn- view of research and educational implications. Exceptional
ing opportunities. Baltimore: Paul H. Brookes. Children, 43, 526-532.
Kamps, D. M., Leonard, B. R., Vernon, S., Dugan, E. P.,
Sugai, G., & White, W. J. (1986). Effects of using object self-
Delquadri, J. C., Gershon, B., Wade, L., & Folk, L. (1992).
stimulation as a reinforcer on the prevocational work rates
Teaching social skills to students with autism to increase
of an autistic child. Journal of Autism and Developmental
peer interactions in an integrated first-grade classroom.
Journal of Applied Behavior Analysis, 25, 281-288. Disorders, 16, 459-471.
Koegel, R. L., Dyer, K., & Bell, L. K. (1987). The influence of Taras, M. E., Matson, J. L., & Leary, C. (1988). Training social
child-preferred activities on autistic children's social behav- interpersonal skills in two autistic children. Journal of Be-
ior. Journal of Applied Behavior Analysis, 20, 243-252. havior Therapy and Experimental Psychiatry, 19, 275-280.
Koegel, R. L., & Egel, A. L. (1979). Motivating autistic chil- Van der Klift, E., & Kunc, N. (1994). Beyond benevolence:
dren. Journal of Applied Behavior Analysis, 88, 418-426. Friendships and the politics of help. In J. S. Thousand, R. A.
Koegel, R. L., & Frea, W. D. (1993). Treatment of social be- Villa, & A. I. Nevin (Eds.), Creativity and collaborative
havior in autism through the modification of pivotal social learning: A practical guide to empowering students and
skills. Journal of Applied Behavior Analysis, 26, 369-377. teachers (pp. 391-401). Baltimore: Paul H. Brookes.
Koegel, R. L., O'Dell, M. C., & Koegel, L. K. (1987). A natu- Williams, T. I. (1989). A social skills group for autistic chil-
ral language teaching paradigm for nonverbal autistic chil- dren. Journal of Applied Behavior Analysis, 19, 143-155.
dren. Journal of Autism and Developmental Disabilities, 17, Wolery, M. R. (1978). Self-stimulatory behavior as a basis for
187-200. devising reinforcers. American Association for the Educa-
Lovaas, I., Newsom, C., & Hickman, C. (1987). Self- tion of the Severely/Profoundly Handicapped Review, 3,23-
stimulatory behavior and perceptual reinforcement. Journal 29.
of Applied Behavior Analysis, 20, 45-68.
Wolery, M. R., Kirk, K., & Gast, D. L. 91985). Self-stimulatory
Odom, S. L., Hoyson, M., Jamieson, B., & Strain, P. S. (1985).
behavior as a reinforcer: Effects and side effects. Journal of
Increasing handicapped preschoolers' peer social interac-
Autism and Developmental Disorders, 15, 149-161.
tions: Cross-setting and component analysis. Journal of Ap-
plied Behavior Analysis, 18, 3-16.
Odom, S. L., & Strain, P. S. (1984). Peer-mediated approaches Received: September 29, 1997
to promoting children's social interaction: A review. Ameri- Final acceptance: April 10, 1998
can Journal of Orthopsychiatry, 54, 544-557. Editor in charge: Mark Durand

Downloaded from rps.sagepub.com at University of Sussex Library on January 9, 2016