You are on page 1of 13

Research in Autism Spectrum Disorders 6 (2012) 1211–1223

Contents lists available at SciVerse ScienceDirect

Research in Autism Spectrum Disorders


Journal homepage: http://ees.elsevier.com/RASD/default.asp

The effect of peer- and sibling-assisted aquatic program on interaction


behaviors and aquatic skills of children with autism spectrum disorders
and their peers/siblings
Chia-Hua Chu, Chien-Yu Pan *
Department of Physical Education, National Kaohsiung Normal University, No. 116, He-Ping First Road, Kaohsiung 802, Taiwan

A R T I C L E I N F O A B S T R A C T

Article history: The purpose of this study was to assess the effect of peer- and sibling-assisted learning on
Received 8 December 2011 interaction behaviors and aquatic skills in children with autism spectrum disorders (ASD).
Received in revised form 6 February 2012 Outcome measures were also examined in their typically developing (TD) peers/siblings.
Accepted 9 February 2012 Twenty-one children with ASD and 21 TD children were assigned in three groups: peer-
assisted (PG), sibling-assisted (SG), and control (CG). All participated in 16-week aquatic
Keywords: settings under three instructional conditions (teacher-directed, peer/sibling-assisted, and
Peer voluntary support). The main findings were that (a) PG and SG of children with ASD
Sibling
showed significantly more improvement on physical and social interactions with their TD
Interaction behaviors
peers/siblings during peer/sibling-assisted condition as compared to CG (p < 0.01), (b) PG
Aquatic skills
Autism
and SG of children with ASD showed significantly more improvement on physical
interactions with their TD peers/siblings (p < 0.01) and social interactions with their
teachers and other children with ASD (p < 0.01) during voluntary support condition as
compared to CG, and (c) all children with ASD and their TD peers/siblings significantly
increased their aquatic skills after the program. The benefit for children with ASD as well as
TD peers/siblings makes the use of TD peer/sibling assisted learning an even more
desirable instructional strategy.
ß 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Children with autism spectrum disorders (ASD) present with difficulties in socialization, communication, and
restricted and repetitive interests and behaviors (American Psychiatric Association, 2000). In addition to these typical
ASD features, there is increasing awareness that some children with ASD experienced considerable difficulties in motor
functions (Pan, Tsai, & Chu, 2009; Provost, Heimerl, & Lopez, 2007; Staples & Reid, 2010). Although there are a wide
variety of treatment strategies, there is no consensus to date as which treatment modality is most effective. It is possible
that typically developing (TD) peer- and sibling-assisted activities provide a more appropriate physical, social and
behavioral environment that will prove beneficial to children with ASD for the development of such skills.
Research has demonstrated that utilizing peers as intervention agents (e.g., peer tutoring, peer-assisted learning) can
be effective in the general education classroom to teach students with disabilities academic gains (Mcduffie,
Mastropieri, & Scruggs, 2009), on-task behaviors (Bowman-Perrott, 2009), and activities of daily living (Gilberts, Agran,
Hughes, & Wehmeyer, 2001). It has also been successful in improving the performances of students with disabilities in

* Corresponding author. Tel.: +886 7 7172930x3531; fax: +886 7 7114633.


E-mail address: chpan@nknucc.nknu.edu.tw (C.-Y. Pan).

1750-9467/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.rasd.2012.02.003
1212 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

the general physical education (PE) classes. For example, it was shown to increase moderate-to-vigorous physical
activity (MVPA) of students who are deaf (Lieberman, Dunn, Mars, & McCubbin, 2000), to improve the correct
performances of motor skills in students with developmental disabilities (Houston-Wilson, Dunn, van der Mars, &
McCubbin, 1997), to improve two tennis strike skills (forehand strike and two-hand side-arm strike) during PE class in
children with a mild intellectual disability (Strickland, Temple, & Walkley, 2005), and to increase total and correct
catches for students with autism (Ward & Ayvazo, 2006). Similarly, DePaepe (1985) studied the Academic Learning
Time-Physical Education (ALT-PE) of students with disabilities in three environments: a self-contained (segregated)
adapted PE class, a mainstreamed PE class, and a class with one-to-one peer tutors. Results indicated that the peer tutor
class yielded significantly more ALT-PE during a static and dynamic balance task than the self-contained or
mainstreamed PE classes, respectively. These studies have occurred in inclusive settings providing empirical support for
peer-mediated intervention as a successful strategy on increasing physical activity levels and motor skills for children
with ASD.
Although not as many studies as peer-mediated intervention advocating its effectiveness, several researchers have
examined the feasibility of using TD siblings in helping children with ASD learn functional skills in natural settings. Taso
and Odom (2006) used a single-subject multiple-baseline design to investigate the effectiveness of a sibling-mediated
intervention in supporting the social behaviors of young children with autism. The researchers taught four sibling dyads
how to socially engage their brothers with autism when playing together. Results revealed that the sibling-mediated
intervention encouraged more social interactions between sibling dyads. Specifically, TD siblings increased their social
initiations toward their brothers with autism, and children with autism engaged in more social behaviors directed
toward their siblings without a disability. Furthermore, children with autism increased their frequency of joint
attention with their TD siblings during the play sessions. Knott, Lewis, and Williams (2007) used naturalistic
observations of sibling pairs including an autistic child and a TD child. Sibling pairs were observed playing with each
other at home, on two occasions, 12 month apart. Results showed that the TD siblings doubled their rate of prosocial
initiations by the second observation. In response to these initiations, there was an increase in the number of positive
responses by children with autism during the first and the second observation. Furthermore, imitation was also
increased during the second observation in children with autism. These sibling-mediated intervention or observation
studies indicated that TD siblings can be effective on promoting social behaviors for children with ASD.
For children with ASD, one area of vital importance is the development of social competence. Since social interaction is a
core deficit associated with ASD, the effectiveness of the interventions aimed at increasing social interactions among
children with ASD have been critically important. However, motor skills performance can have significant long-term effects
on a variety of emotional, social and behavioral development. Past research has found that children and adolescents with
poor motor abilities had higher levels of anxiety and lower levels of perceived self-worth (Skinner & Piek, 2001), peer
relations (Livesey, Mow, Toshack, & Zheng, 2011), and scholastic ability and athletic competence (Piek, Baynam, & Barrett,
2006). Furthermore, poor motor skills can cause decreased participation in physical activity (Cairney, Hay, Faught, Corna, &
Flouris, 2006) and below-average performance on different components of physical fitness (Cantell, Crawford, & Doyle-
Parker, 2008), and in the long run, put their health at risk. Therefore, physical activity intervention should not be overlooked
in this population. There is a need to identify modes or types of physical activity intervention to improve both motor skills
and social behaviors with these children.
Attwood (1998) suggested that the ability to swim was less affected in children with ASD than other physical
activities because swimming can enhance a child’s competence and foster an appreciation of proficient movement.
Buoyancy enables initiation of independent movement possibilities that are difficult to achieve on land due to
gravitational restraints (Hutzler, Chacham, Bergman, & Szeinberg, 1998). One latest aquatic intervention with a social
skills component for children with ASD has been reported, although not specific to peer- or sibling-assisted aquatic
program. Pan (2010) used a controlled, single-blinded design to evaluate the effect of a 10-week water exercise
swimming program on the aquatic skills and social behaviors in 16 children with ASD. The water exercise swimming
program was designed based on the Humphries’ Assessment of Aquatic Readiness (HAAR, Humphries, 2008) instruction
and utilized the Treatment of Education of Autistic and related Communication Handicapped Children (TEACCH) model
(Blubaugh & Kohlmann, 2006). The aquatic skills assessment was performed by the instructors and social behaviors
were rated by the child’s school classroom teacher before and after 10 weeks training. The result of this study (Pan,
2010) indicated that the 10-week water exercise swimming program improved the child’s aquatic skills and social
behaviors after the intervention, suggesting that the inclusion of physical activity treatment components targeting
social skills could be made.
The aim of this study was to examine the effects of peer- and sibling-assisted aquatic program on the interaction
behaviors and aquatic skills in children with ASD. Additional variable of interest was interaction behaviors and aquatic
skills presented by TD peers and siblings. This study was conducted in an after-school aquatic program containing three
instructional conditions regarding support provided to children with ASD across teacher-directed, peer/sibling-assisted,
and voluntary peer/sibling support conditions. It was hypothesized that (a) the participant children with ASD would
increase interaction behaviors (physical and social interactions) with their trained TD peers/siblings during peer/
sibling-assisted and voluntary peer/sibling support conditions, (b) trained TD peers/siblings would increase interaction
behaviors with children with ASD during peer/sibling-assisted and voluntary peer/sibling support conditions, and (c)
both trained TD peers/siblings and children with ASD would increase their aquatic skills after the program.
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223 1213

2. Methods

2.1. Participants

Participants for this study included 21 children with ASD and 21 TD children from middle income families that lived in
urban communities in the southern region of Taiwan. They were recruited through posted advertisements in the community,
schools, social/service organizations, and the Autism Society of Taiwan. The sampling design was purposive, meaning criteria
were used to select children who were representative of children with ASD, TD siblings of children with ASD, and TD children
who might be trained to be peers. All participants were then assigned in three groups: peer-assisted group (PG: ASD, n = 7;
TD, n = 7), sibling-assisted group (SG: ASD, n = 7; TD, n = 7), and control group (CG: ASD, n = 7; TD, n = 7).
Participants with ASD were diagnosed based on the clinical diagnosis according to the Diagnostic and Statistical Manual
of Mental Disorders (DSM-IV) criteria for autistic disorder (American Psychiatric Association, 2000) by certified child
psychiatrists specialized in clinical service of children with ASD (Taiwan Executive Yuan Department of Health, 2010).
Children with ASD included 14 mild or high-functioning autism and 7 Asperger’s syndrome. Level of severity (mild,
moderate, severe, and very severe) was based on functioning in the social adaptive skill areas and language comprehension
and expression (Taiwan Executive Yuan Council of Labor Affairs, 2010). No participant with ASD had a history of previous or
current psychiatric or neurological disorder aside from ASD. At the time of the study, no participant with ASD received
prescribed medications. All participants with ASD attended inclusive schools and were assigned to the resource room on a
regularly scheduled basis to receive special education services while continuing their other studies in regular classrooms
during most of each school day. None were enrolled in an inclusive community-based physical activity program.
TD participants were recruited from similar neighborhoods to the participants with ASD. They were eligible to participate
in the study if they had no history or ongoing medical treatment and met the inclusion criteria for each group. For SG, both the
child with ASD and his/her TD sibling were recruited if they (a) were healthy, (b) were between ages 7 and 12 years old, (c)
lived together, and (d) were able to follow instructions and the requirements of the sibling-assisted aquatic program. For PG
and CG of TD children, they were recruited if they were (a) healthy, (b) within 7–12 months of age as the other participants had
to be, and (c) able to follow the requirements of either the peer-assisted aquatic program or control instructions.
With regard to prior swim experience, all 42 participants had no prior experience or training in aquatic skills. They had
very limited breathing, floating, and stroke skills based on parent reports and first assessment at study entry. The majority of
participants had accomplished mental adjustment stage (stage I) with no fear of the water. The university human subjects
review board approved this study prior to the participant recruitment. Written informed consent forms were obtained from
all parents and children prior to enrollment in the study. All children completed the whole 16-week program. Descriptive
characteristics are found in Table 1.

2.2. Peer- and sibling-assisted training

PG and SG of TD peers and siblings were trained individually by the primary researcher before this study began,
using the steps and methods proposed by Klavina and Block (2008). At the beginning, the primary researcher and PG as
well as SG of peers/siblings discussed differences in people, the way they would like to be treated, and the way to treat
children with ASD. Next, the primary researcher and these peers/siblings discussed rules and roles of being a designated
peer/sibling (e.g., being friendly, talk softly, and providing praises). Then, the assisting steps were introduced: (a)
physical interactions (e.g., cues, prompts, demonstration, physical assistance, feedback (e.g., general, specific)) and (b)
social interactions (e.g., interdependent play, mutual organization, gestures, and talking together). The primary
researcher conducted four 30-min training sessions for these peers/siblings. For the first two 30-min training sessions,
these peers/siblings were trained without the children with ASD. On the third and fourth sessions, these peers/siblings
and children with ASD were assigned to work in pairs to practice the assisting steps. Selected activities matched the
aquatic program and allowed implementing training practice appropriate to the real aquatic program situation. All
activities were monitored by the primary researcher to ensure that each trained peer/sibling had an opportunity to
practice with the target child with ASD. Once PG and SG of TD peers and siblings were trained, they received ongoing
training for 5–10 min per class throughout the program.
To determine the effectiveness of the training program, trained peers/siblings took written and practical exams
covering physical and social interaction behaviors. The practical exam was assessed by the primary researcher for each
trained peer/sibling when he or she was paired with the child with ASD. For example, trained peers/siblings were given
the following instructions: ‘‘During the peer/sibling-assisted condition, you will have to observe your peer with ASD and
give him/her positive, pleasant, and reinforcing responses.’’ ‘‘During the voluntary peer/sibling support condition, you
will be free to take care of your peer with ASD with the specified games, and to give him/her advice until the next
round.’’ These peers/siblings practiced, in turn, demonstrating, providing feedback and encouragement using assisting
steps introduced in the training. Other examples of scenarios included: ‘‘What would you respond if the child with ASD
exhibited off-task behavior?’’ ‘‘What would you do if the child with ASD needed encouragement during cooperative
games?’’ The trained peer/sibling had to score 90% or better on the knowledge and practical exam. If the peer/sibling
received less than 90% on the written and the practice, then additional training sessions were added until this standard
was met.
1214 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

Table 1
Descriptive statistics of participants.

Peer-assisted (n = 14) Sibling-assisted (n = 14) Control (n = 14) x2 p

ASD (n = 7) TD Peer (n = 7) ASD (n = 7) TD Sibling ASD (n = 7) TD Control


(n = 7) (n = 7)

M SD M SD M SD M SD M SD M SD

Age (years) 8.54 1.62 7.18 1.07 8.33 1.56 7.33 2.41 9.28 1.46 8.69 2.88 8.51 0.13
Height (cm) 125.76 11.91 124.43 5.38 132.20 11.28 123.79 15.35 135.70 6.17 133.33 24.06 6.89 0.23
Weight (kg) 25.60 4.91 25.04 5.60 34.47 13.11 30.00 13.57 33.64 8.35 33.80 15.26 7.58 0.18
BMI (kg/m2) 16.16 1.60 16.13 3.45 19.41 5.52 18.74 4.01 18.43 3.42 18.20 2.31 8.04 0.15
Gender
Female 1 6 0 5 0 4
Male 6 1 7 2 7 3
Therapy
PT 2 – 4 – 3 –
GT 0 – 1 – 1 –
OT 1 – 1 – 0 –
ST 0 – 1 – 0 –

Note. ASD, autism spectrum disorders; TD, typically developing; BMI, body mass index; PT, physical therapy; GT, group therapy; OT, occupational therapy;
ST, speech therapy.

2.3. Setting

Before the study began, a workshop was held to introduce the aquatic program to both PG and SG of TD children. Aquatic
instructors in PG and SG had also completed a training course conducted by the primary researcher prior to the study. This
training was on how to interact with ASD children in the aquatic program. All three groups of instructors had an
undergraduate degree in PE and had 1–2 years direct experience working with individuals with ASD and their needs. During
each aquatic program session, two paired children (one ASD/TD dyad) were with the same instructor for each session.
The after-school aquatic program was 60 min in length and held twice per week for 16 weeks at a local indoor
hydrotherapy and swimming pool in Kaohsiung city, Taiwan. The program consisted 10 min of the warm-up activities,
35 min of the one-to-two teacher-to-students instruction (20 min and 15 min of the teacher-directed and the peer/sibling-
assisted conditions, respectively), and 15 min of group games (voluntary peer/sibling support condition) and cool down
activities at the end of class. All three groups of children participated in all three conditions, but CG was simply a group of
children who were only instructed by the instructors without providing any specific peer-assisted interactions.
Within teacher-directed (e.g., baseline) instructional conditions, the instructor provided instructions to both the child
with ASD and TD child. The TD child was not instructed on how to assist the child with ASD during these conditions.
Therefore, data were obtained from intact aquatic sessions. The goal was to ensure that both children had adequately learned
the appropriate aquatic skills.
Immediately following teacher-directed conditions, the instructor in PG and SG initiated experimental (peer/sibling-
assisted) procedures and each trained peer/sibling interacted with the child with ASD. Each CG of TD peers during these
conditions was told to interact with his/her child with ASD in a typical manner. Because TD children in PG and SG had very
limited aquatic skills, they simply demonstrated the skills just learned, physically assisted and gave cues or feedback to the
child with ASD on his or her performance during peer/sibling-assisted conditions. The instructor monitored assisting
activities next to both children and prompted the peer/sibling to interact with the child with ASD if no interactions were
occurring for about 20 s. The instructor praised both the peer/sibling and the child with ASD on successful partnership and
collaboration to help maintain children’s confidence and enjoyment. The primary researcher met the instructor and the
trained peers/siblings prior to each class to give them information on various teaching and interacting techniques in regard
to the day’s activity. Ongoing feedback was provided to the trained peers/siblings after each class to correct physical and
social interaction behaviors and improve the ways they provided these interactions (adapted from Klavina & Block, 2008).
Throughout the peer/sibling-assisted conditions, the primary researcher monitored and gave feedback to the instructors and
the trained peers/siblings when necessary.
The period of voluntary peer/sibling support condition was designated for collaborative games or other activities
encouraging spontaneous and voluntary interactions between children. Children usually engaged in pairs or small group to
maintain high level of interactions and participations. After the period of assigned peer/sibling’s assistance, the instructor
provided these trained peers/siblings an option to choose whether they would be willing to assist their assigned children
with ASD or other children with ASD. If any trained peers/siblings agreed to continue assisting, the instructor let him or her to
do so. Otherwise, the instructor assisted the child with ASD. In addition, untrained TD peers in CG could join or assist children
with ASD. As a result, this condition focused on determining if all TD children (trained peers/siblings and others) would
initiate and engage in voluntary and spontaneous interactions with children with ASD.
The aquatic program was designed based on the Humphries’ Assessment of Aquatic Readiness (HAAR, Humphries, 2008)
assessment instruction developed according to the foundations of the Halliwick Method (Martin, 1981). This method utilizes
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223 1215

the natural progression of the way humans acquire physical movements, and these skills were broken down into several
enabling activities and were taught in a sequential order with respect to children’s acquisition. In the present study, children
were taught on a one-to-two ratio of instructor to student until complete independence was achieved. The student–
instructor pair becomes a unit within a group activity so that the student gains the advantages of social interaction with his/
her peers while at the same time enjoying the constant attention of an individual instructor. The ultimate goal was to
increase aquatic and social skills of both children with ASD and TD children. All participants had 16 weeks of the prescribed
individual-paced aquatic activity during teacher-directed condition each session. Structured teaching, the most crucial and
recognizable feature of the TEACCH model (Mesibov, Shea, & Schopler, 2004), was also utilized in the program including
organization of the physical environment (e.g., establish clear boundary markings to help children know where they may and
may not go) and visual schedules and work systems (e.g., a board with pictures and words to describe the routine and the
daily aquatic activity). The aquatic program was implemented as prescribed and supervised by the primary researcher each
session.

2.4. Measurements of dependent variables

2.4.1. Interaction behaviors


The operational measurements in two main behavior categories (physical and social) are classified as interaction
behaviors according to the Computerized Evaluation Protocol of Interactions in Physical Education (CEPI-PE, Klavina &
Selavo, 2006). The physical interaction behaviors encompassed one-to-one interactions related to the aquatic activities,
and the social interaction behaviors included verbal or nonverbal communication on content not related to aquatic
program. While these behaviors were not related to the program, they might affect social acceptance and relationships
between children with ASD and their paired children without a disability. Four subcategories in each main interaction
behavior distinguished interactions between (a) the child with ASD and the instructor, (b) the child with ASD and the
designated peer/sibling, (c) the child with ASD and an other peer/sibling, and (d) the child with ASD and an other
children with ASD. While this study primary focused on examining changes in interaction behaviors of children with
ASD, the additional variable of interest also was behaviors presented by peers/siblings.
In the current study, physical interaction behaviors were active engagement in individual or group activities related to
objective or tasks in aquatic program when participants were learning or practicing on a physical activity (e.g., hula-hoop
swimming) or a skill (e.g., glide with straight leg kick). Four tasks were included: activity engagement with assistance from a
trained peer or sibling (Ap/As); activity engagement with assistance from a teacher (At); activity engagement with assistance
from an other peer or sibling (Aop/Aos); activity engagement together with a child with ASD (Aa).
Social interaction behaviors were voluntary and spontaneous exchange of observable responses between two children,
which were not academic or instructional. In the current study, these behaviors were positive (e.g., pleasant and reinforcing
responses) and could be verbal (e.g., said ‘‘hi’’) or nonverbal (e.g., gave a hug). Four tasks were also included: social
interactions with a trained peer or sibling (Sp/Ss); social interactions with a teacher (St); social interactions with an other
peer or sibling (Sop/Sos); social interactions with a child with ASD (Sa).
The CEPI-PE categorical variables have been validated in two pilot studies for its use with elementary school students in
general PE setting (Klavina, 2007). The instrument evidences a high degree of inter- and intra-observer reliability. It also
demonstrates face validity and concurrent validity. The interaction behaviors can be observed and coded using interval or
continuous recording techniques. In this study a five-second observation and five-second record partial interval system was
applied. The obtained data were analyzed based on percentage mean scores.

2.4.2. Aquatic skills


The HAAR checklist was used to assess each participant’s aquatic skills. It is designed based on the Halliwick Method
(Martin, 1981) which has been reported to be safe for people of all ages and with many types of disabilities (MacKinnon,
2003). The HAAR demonstrates face validity and a high degree of inter-rater reliability (above 0.90, Humphries, 2008). It is
divided into five stages: (a) mental adjustment (5 tasks; e.g., will enter/exit the pool without holding the instructor’s hand),
(b) introduction to water environment (10 tasks; e.g., will blow bubbles in water), (c) rotations (3 tasks; e.g., will roll from
supine position to prone position and back to supine), (d) balance and control (8 tasks; e.g., will move arms in rudimentary
pattern with instructor’s support), and (e) independent movement in water (6 tasks; e.g., will come to instructor with
straight leg kick and straight over arm independently). Tasks in each stage of HAAR can be found in Humphries (2008) or Pan
(2010).
For each task that the participant was able to complete he/she was given a score of ‘‘1.’’ If the participant was unable to
complete the task he/she was given a score of ‘‘0.’’ At the end of the assessment, the participant’s total score for each stage
was added up. A percentage score of each stage was used for data analysis. It was calculated by dividing the participant’s total
score for that stage by the total possible points for that stage, multiplied by 100.

2.5. Data collection and procedure

The administration aquatic skill test procedures were as follows: each child was in the water with his/her instructor,
while the primary researcher was out of the water and near by. The primary researcher told the instructor the nature of each
1216 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

skill and the child was assessed according to his/her performance of each specific skill. Inter-observer reliability data have
been collected on each of the 42 participants before and after the program by two trained graduate students and the primary
researcher. It was calculated by dividing agreements by agreements plus disagreements and multiplying by 100 for each
stage. Results showed a high level of inter-observer agreement (all above 0.90).
In this study, interaction behaviors data were collected for a total of 32 sessions. The length of each observation
session was 50 min divided into three parts: (a) 20 min of teacher-directed instructional condition, (b) 15 min of the
peer/sibling-assisted instructional condition, and (c) 15 min of voluntary peer/sibling support condition. All observation
sessions were collected on videotapes with the use of the SONY Mini DV Digital Handycam (HDR-SR12) video recorder.

2.5.1. Observer training


All 32 lessons were videotaped and coded. Two observers received training prior to the start of data collection that
included a copy of definitions for all interaction behaviors, explanations about each behavior, and its critical elements.
Observers then practiced coding from videotapes and received feedback. Finally, observers were assessed on their coding
accuracy using a 30-min tape and were required to meet a criterion level of 100% accuracy.
Reliability checks were completed randomly for three conditions on each of the behaviors by two trained graduate students
and the primary researcher. Two graduate students rated frequency of behavior for all sessions, and the primary researcher rated
33% of sessions for each behavior. The reliability was calculated by dividing agreements by agreements plus disagreements and
multiplying by 100 for each condition. Results showed a high level of inter-observer agreement (all above 0.90).

2.6. Procedural reliability and social validity

In order to ensure that quality peer/sibling assistance took place, the aquatic program was implemented as prescribed
and supervised by the primary researcher each session. If trained peer/sibling was off task in the class, the instructor
managed the behavior, and if trained peer/sibling was off task within the peer/sibling assisted situation, the primary
researcher intervened to assist trained peer/sibling. In addition, each instructor completed a self-report checklist of their
compliance with the teaching steps after each session to assess whether the peer/sibling assisted instruction was
implemented with precision and consistency across participants. By the end of the study, the instructors of PG and SG had
completed 100% of the checklist forms, and all of them responded ‘‘yes’’ to the steps across each peer/sibling-assisted session.
Social validation data were obtained based on three sources. First, instructors involved in PG and SG completed a brief
questionnaire with the purpose to obtain their opinions about the goal, procedures, and outcomes of the peer/sibling assisted
condition as well as assisting strategies that might be used in future studies. The questionnaire consisted of 2 open-ended
questions and 13 Likert scale items (rated on a 7-point scale) questions. Second, at the end of this study, trained peers/
siblings were asked on how they felt about their tasks and outcomes of assisting procedures. Finally, anecdotal social validity
information (e.g., conversations, correspondence) was collected informally from parents when the primary researcher was
contacted by parents to discuss the aquatic program.

2.7. Statistical analysis

Descriptive analyses were conducted for each stage of aquatic skills and each condition of interaction behaviors in children
with ASD and TD children. Due to small sample size, the nonparametric Wilcoxon signed ranks tests were chosen for pre-post
test pairwise comparisons for aquatic skills within each group. The Kruskal Wallis tests were performed to compare aquatic skills
between pre- and post-test scores among groups. Differences in interaction behaviors for each condition among groups were
also analyzed via the nonparametric Kruskal Wallis tests. Effects sizes (Cohen’s d) were computed to quantify the magnitude of
changes in interaction behaviors among three conditions within groups. These were calculated by dividing the mean change in a
condition score by the standard deviation of the condition score at baseline to quantify the magnitude of change between two
condition scores. For example, for PG, score differences in the interaction behaviors between teacher-directed and peer-assisted
conditions and between peer-assisted and voluntary support conditions indicated effectiveness between peer-assisted
condition and regular teacher-directed condition and carry-over effects of the peer-assisted condition, respectively. All
statistical analyses were conducted with SPSS version 13.0 for windows, and statistical significance was set at p < 0.05.

3. Results

Preliminary data analyses were conducted to determine the possibility of gender effect on demographic variables. No
gender effect (age, x2 = 2.14, p = 0.14; height, x2 = 2.66, p = 0.10; weight, x2 = 3.39, p = 0.07; BMI, x2 = 1.13, p = 0.29) was
found from the Kruskal Wallis tests, and therefore, gender was not used in subsequent analyses.

3.1. Aquatic skills for children with ASD and TD children

Results for aquatic skills are found in Table 2. The Kruskal Wallis tests showed that there were no statistically significant
differences between groups at study entry and after the program. For all children with ASD and TD children, a statistically
significant improvement of score after the program was found in stages II, III, IV, and V (all p < 0.05).
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223
Table 2
Summary of the nonparametric Wilcoxon signed ranks tests on the aquatic skills between pre- and post-test scores in each group.

Stage Peer-assisted (n = 14) Sibling-assisted (n = 14) Control (n = 14)

ASD (n = 7) TD Peer (n = 7) ASD (n = 7) TD Sibling (n = 7) ASD (n = 7) TD Control (n = 7)


a b a b a b a b a b
Pre Post Z Pre Post Z Pre Post Z Pre Post Z Pre Post Z Prea Postb Z

I 97.14 100.00 1.00 94.29 100.00 1.00 94.29 100.00 1.41 91.43 100.00 1.34 97.14 100.00 1.00 94.29 100.00 1.41
(7.56) (0.00) (15.12) (0.00) (9.76) (0.00) (15.74) (0.00) (7.56) (0.00) (9.76) (0.00)
II 64.29 94.29 2.23* 68.57 100.00 2.38* 61.43 100.00 2.39* 72.86 100.00 2.41* 52.86 100.00 2.40* 75.71 98.57 2.46*
(24.40) (15.12) (17.73) (0.00) (14.64) (0.00) (11.13) (0.00) (12.54) (0.00) (5.35) (3.78)
III 9.52 42.85 2.65** 4.76 38.09 2.33* 9.52 38.09 2.12* 14.28 61.90 2.43* 28.57 71.43 2.04* 14.28 47.62 2.06*
(25.20) (25.20) (12.60) (12.60) (25.20) (23.00) (17.82) (35.64) (23.00) (35.64) (17.82) (26.23)
VI 32.14 83.93 2.23* 28.57 100.00 2.39* 37.50 94.64 2.45* 50.00 94.64 2.41* 33.93 100.00 2.40* 44.64 92.86 2.39*
(33.74) (18.70) (21.30) (0.00) (12.50) (9.83) (19.09) (9.83) (15.67) (0.00) (6.68) (12.20)
V 14.29 59.52 2.03* 7.14 80.95 2.39* 26.19 78.57 2.39* 19.05 80.95 2.39* 38.09 88.09 2.26* 11.90 69.05 2.43*
(37.80) (34.50) (18.90) (11.50) (39.51) (20.89) (14.99) (27.94) (34.31) (18.55) (15.85) (22.42)

Note. M(SD); ASD, autism spectrum disorders; TD, typically developing.


* p < 0.05.
** p < 0.01.
a
Kruskal Wallis tests for pre-test between groups: stage I, x2 = 1.17, p = 0.95; stage II, x2 = 9.62, p = 0.09; stage III, x2 = 7.12, p = 0.21; stage VI, x2 = 6.83, p = 0.23; stage V, x2 = 7.32, p = 0.20.
b
Kruskal Wallis tests for post-test between groups: stage I, x2 = 0.00, p = 1.00; stage II, x2 = 4.10, p = 0.54; stage III, x2 = 6.36, p = 0.27; stage VI, x2 = 9.03, p = 0.11; stage V, x2 = 5.41, p = 0.37.

1217
1218 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

Table 3
Summary of the nonparametric Kruskal Wallis tests on the teacher-directed condition for each group.

ASD (n = 21) x2 p TD (n = 21) x2 p

PG (n = 7) SG (n = 7) CG (n = 7) PG (n = 7) SG (n = 7) CG (n = 7)

Physical
Ap/As 0.05 (0.03) 0.03 (0.04) 0.02 (0.04) 3.49 0.18 – – – – –
At 46.79 (6.69) 52.69 (14.01) 53.98 (6.69) 4.44 0.11 51.38 (9.91) 52.76 (6.09) 45.17 (13.38) 1.61 0.45
Aop/Aos 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 1.00 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 1.00
Aa 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 1.00 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 1.00
Social
Sp/Ss 0.27 (0.26) 0.51 (0.94) 0.11 (0.11) 2.02 0.36 – – – – –
St 0.13 (0.06) 0.18 (0.36) 0.13 (0.17) 1.99 0.37 1.00 (0.93) 1.16 (2.10) 0.24 (0.36) 4.49 0.11
Sop/Sos 0.01 (0.01) 0.01 (0.02) 0.00 (0.00) 3.60 0.17 0.08 (0.12) 0.01 (0.02) 0.01 (0.02) 4.79 0.09
Sa 0.03 (0.04) 0.01 (0.02) 0.00 (0.01) 2.12 0.35 0.39 (0.53) 0.63 (1.15) 0.16 (0.18) 0.28 0.87

Note. M(SD); PG, peer-assisted group; SG, sibling-assisted group; CG, control group; Ap/As, activity engagement with assistance from a trained peer or
sibling; At, activity engagement with assistance from a teacher; Aop/Aos, activity engagement with assistance from an other peer or sibling; Aa, activity
engagement together with a child with ASD; Sp/Ss, social interactions with a trained peer or sibling; St, social interactions with a teacher; Sop/Sos, social
interactions with an other peer or sibling; Sa, social interactions with a child with ASD.

3.2. Interaction behaviors during teacher-directed condition between groups

The mean and standard deviation of subcategories of interaction behaviors for children with ASD and TD children during
teacher-directed condition are illustrated in Table 3. These data from the Kruskal Wallis tests show that both peer- and
sibling-assisted children with ASD did not differ statistically on any of the interaction behaviors from those of children in CG.
Physical and social interaction behaviors were also similar between trained peers/siblings and untrained TD children.

3.3. Interaction behaviors during peer/sibling-assisted condition between groups

3.3.1. Children with ASD


Results from the Kruskal Wallis tests (Table 4) showed that activity engagement with assistance from the designated TD
children were significantly greater for both PG and SG as compared to CG (x2 = 11.86, p < 0.01). CG of children with ASD had
significantly higher activity engagement with assistance from their teachers as compared to both PG and SG of children with
ASD (x2 = 6.43, p < 0.05). Significant differences on social interaction with the designated TD children and teachers were also
found among groups: both PG and SG of children with ASD demonstrated significantly higher social interaction with their
peers/siblings (x2 = 8.55, p < 0.05) and teachers (x2 = 6.59, p < 0.05) than those of children in CG.

3.3.2. TD children
As shown in Table 4, both PG and SG of TD children demonstrated significantly more social interaction with another child
with ASD than those of children in CG (x2 = 6.51, p < 0.05). None of significant differences on any of the physical interaction
subcategories were found among groups.

Table 4
Summary of the nonparametric Kruskal Wallis tests on the peer- or sibling-assisted condition for each group.

ASD (n = 21) x2 Post hoc TD (n = 21) x2 Post hoc

1. PG (n = 7) 2. SG (n = 7) 3. CG (n = 7) 4. PG (n = 7) 5. SG (n = 7) 6. CG (n = 7)

Physical
Ap/As 12.04 (11.27) 12.87 (5.83) 0.03 (0.06) 11.86** 1, 2 > 3 – – – –
At 44.74 (10.18) 43.64 (2.83) 53.01 (5.23) 6.43* 1, 2 < 3 43.04 (8.50) 45.16 (5.05) 48.34 (14.00) 1.25
Aop/Aos 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00
Aa 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00
Social
Sp/Ss 0.95 (0.64) 1.37 (1.57) 0.13 (0.12) 8.55* 1, 2 > 3 – – – –
St 1.54 (1.39) 1.64 (2.20) 0.16 (0.15) 6.59* 1, 2 > 3 0.99 (0.97) 1.21 (1.71) 0.29 (0.39) 3.59
Sop/Sos 0.01 (0.01) 0.01 (0.02) 0.00 (0.00) 3.60 0.07 (0.07) 0.01 (0.02) 0.06 (0.13) 2.97
Sa 0.14 (0.21) 0.13 (0.34) 0.01 (0.02) 5.83 0.81 (0.40) 1.28 (1.46) 0.22 (0.27) 6.51* 4, 5 > 6

Note. M(SD); PG, peer-assisted group; SG, sibling-assisted group; CG, control group; Ap/As, activity engagement with assistance from a trained peer or
sibling; At, activity engagement with assistance from a teacher; Aop/Aos, activity engagement with assistance from an other peer or sibling; Aa, activity
engagement together with a child with ASD; Sp/Ss, social interactions with a trained peer or sibling; St, social interactions with a teacher; Sop/Sos, social
interactions with an other peer or sibling; Sa, social interactions with a child with ASD.
* p < 0.05.
** p < 0.01.
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223 1219

Table 5
Summary of the nonparametric Kruskal Wallis tests on the voluntary peer/sibling support condition for each group.

ASD (n = 21) x2 Post hoc TD (n = 21) x2 Post hoc

1. PG (n = 7) 2. SG (n = 7) 3. CG (n = 7) 4. PG (n = 7) 5. SG (n = 7) 6. CG (n = 7)

Physical
Ap/As 4.16 (1.93) 4.65 (2.62) 0.09 (0.07) 13.37** 1, 2 > 3 – – – –
At 35.03 (10.56) 37.00 (5.17) 23.96 (5.38) 11.27** 1, 2 > 3 25.80 (6.67) 30.75 (4.38) 29.92 (13.33) 1.40
Aop/Aos 2.46 (2.21) 1.96 (0.88) 1.22 (0.64) 2.76 3.67 (1.87) 3.29 (2.02) 2.77 (2.30) 0.50
Aa 3.55 (1.33) 3.95 (3.11) 3.37 (1.83) 0.01 6.46 (1.12) 6.83 (2.61) 2.67 (1.52) 10.78** 4, 5 > 6
Social
Sp/Ss 2.43 (2.21) 2.45 (2.30) 1.21 (0.82) 1.16 – – – –
St 2.67 (2.32) 3.28 (1.92) 0.43 (0.51) 12.26** 1, 2 > 3 2.51 (1.40) 2.34 (0.56) 0.48 (0.58) 10.81** 4, 5 > 6
Sop/Sos 0.77 (0.91) 0.46 (0.34) 0.30 (0.27) 1.35 3.11 (2.06) 2.04 (3.30) 1.31 (2.05) 2.81
Sa 1.62 (1.32) 1.50 (0.26) 0.01 (0.01) 13.77** 1, 2 > 3 3.46 (3.78) 2.89 (2.51) 1.57 (1.17) 0.62

Note. M(SD); PG, peer-assisted group; SG, sibling-assisted group; CG, control group; Ap/As, activity engagement with assistance from a trained peer or
sibling; At, activity engagement with assistance from a teacher; Aop/Aos, activity engagement with assistance from an other peer or sibling; Aa, activity
engagement together with a child with ASD; Sp/Ss, social interactions with a trained peer or sibling; St, social interactions with a teacher; Sop/Sos, social
interactions with an other peer or sibling; Sa, social interactions with a child with ASD.
** p < 0.01.

3.4. Interaction behaviors during voluntary peer/sibling support condition between groups

3.4.1. Children with ASD


The Kruskal Wallis test results (Table 5) showed that both PG and SG of children with ASD had significantly higher
physical interaction behavior (activity engagement with assistance from a trained peer or sibling) than those of children
in CG (x2 = 13.37, p < 0.01). Both PG and SG of children with ASD also demonstrated significantly more activity
engagement with assistance from a teacher than those of children in CG (x2 = 11.27, p < 0.01). For the social interaction
behaviors, children with ASD in PG and SG had significantly greater social interactions with teachers (x2 = 12.26,
p < 0.01) and another child with ASD (x2 = 13.77, p < 0.01) as compared to those of children in CG.

3.4.2. TD children
Activity engagement together with a child with ASD was found to be significantly higher for both PG and SG of trained
peers and siblings than untrained peers in CG (x2 = 10.78, p < 0.01). Both trained peers and siblings also demonstrated more
social interactions with a teacher than untrained peers (x2 = 10.81, p < 0.01).

3.5. Differences on interaction behaviors among three conditions within group of children with ASD

Estimates of effect size (ES) on interaction behaviors for children with ASD among three conditions are summarized
in Table 6. For PG, children with ASD had significant improvements (large ESs, all Cohen’s d 3 0.8) on physical

Table 6
Summary of effect sizes on physical/activity and social interaction behaviors for children with ASD.

Teacher-directed condition – Peer/Sibling assisted condition Peer/Sibling assisted condition – Voluntary support condition

1. Peer-assisted 2. Sibling-assisted 3. Control 1. Peer-assisted 2. Sibling-assisted 3. Control

Change ES Change ES Change ES Change ES Change ES Change ES

Physical/activity interaction
Ap/As 11.99 348.82a 12.84 302.94a 0.01 0.17 7.88 0.70b 8.22 1.41a 0.06 1.09a
At 2.05 0.31c 2.89 0.21c 0.97 0.15 9.71 0.95a 12.80 1.66a 29.05 5.56a
Aop/Aos 0.00 – 0.00 – 0.00 – 2.46 – 1.96 – 1.22 –
Aa 0.00 – 0.00 – 0.00 – 3.55 – 3.95 – 3.37 –
Social interaction
Sp/Ss 0.68 2.62a 0.85 0.91a 0.02 0.17 1.49 2.32b 1.08 0.69b 1.08 8.94a
St 1.41 23.03a 1.46 4.05a 0.03 0.19 1.14 0.82a 1.64 0.75b 0.28 1.90a
Sop/Sos 0.00 0.00 0.00 0.09 0.00 – 0.76 79.75a 0.45 23.81a 0.30 –
Sa 0.12 3.03a 0.12 5.29a 0.00 0.13 1.48 6.94a 1.37 4.03a 0.00 0.19

Note. Ap/As, activity engagement with assistance from a trained peer or sibling; At, activity engagement with assistance from a teacher; Aop/Aos, activity
engagement with assistance from an other peer or sibling; Aa, activity engagement together with a child with ASD; Sp/Ss, social interactions with a trained
peer or sibling; St, social interactions with a teacher; Sop/Sos, social interactions with an other peer or sibling; Sa, social interactions with a child with ASD;
ES, effect size (Cohen’s d).
Positive values indicated improvements and negative values indicated reductions in interaction behaviors.
a
A Cohen’s d value 3 0.8 indicates a large effects size.
b
A Cohen’s d value 3 0.5 < 0.8 indicates a medium effects size.
c
A Cohen’s d value 3 0.2 < 0.5 indicates a small effects size.
1220 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

interactions with their trained peers and all social interaction subcategories except social interactions with the peers at
the peer-assisted condition. Small ES (Cohen’s d value 3 0.2 < 0.5) for the reductions in physical interactions with the
teachers was also noted at the peer-assisted condition. At the voluntary support condition, children with ASD had
significant improvements (three large and one medium ESs [Cohen’s d value 3 0.2 < 0.5]) on all social interaction
subcategories. Reductions in physical interactions with peers/siblings (medium ES) and teachers (large ES) were also
found. Medium ES existed in physical and social interactions with peers/siblings, and large ESs existed for physical and
social interactions with teachers as well as social interactions with other peers/siblings and children with ASD.
SG of children with ASD demonstrated significant improvements on physical interactions with their siblings and social
interactions with their siblings and teachers as well as other children with ASD at the sibling-assisted condition. Physical
interactions with the teachers were decreased at this condition. Large ESs were found for all social interaction subcategories
and physical interactions with siblings. Small ES was found for physical interactions with the teachers. At the voluntary
support condition, all social interaction subcategories increased and two physical interaction subcategories decreased in
sibling-assisted group of children with ASD. Large ESs were noted for physical interactions with peers/siblings and teachers
as well as social interactions with other peers/siblings and children with ASD. Medium ESs were found for social interactions
with peers/siblings and teachers.
CG of children with ASD did not show any significant improvements or reductions on all physical and social
interaction subcategories at the peer/sibling-assisted condition. At the voluntary support condition, CG of children with
ASD showed significantly improvements on physical and social interactions with peers/siblings and social interactions
with teachers. Physical interactions with teachers were significantly decreased. Large ESs were found for all these
changes.

3.6. Social validity

All instructors agreed that improvement in interaction behaviors and aquatic skills were important goals. Instructors
indicated that peer/sibling-assisted condition was a good idea, and it was feasible to implement. They responded that
interactions between trained peers/siblings and children with ASD increased during peer/sibling-assisted and voluntary
peer/sibling support conditions. They also reported observable social growth in these TD children and positive acceptance
toward children with ASD.
Overall, trained peers and siblings were positive about children with ASD and peer/sibling-assisted strategy. They felt that
they enjoyed using peer/sibling assistance and rated physical performance of their children with ASD as ‘‘very good,’’ two as
‘‘good,’’ and one as ‘‘moderate.’’ All trained peers/siblings rated their children with ASD as well liked as other TD peers in their
school classes. Most trained peers/siblings mentioned that getting to know their children with ASD and helping him or her
participate in aquatic activities was a good experience in this study. One trained peer expressed that her least favorite part of
peer assisting was that she occasionally had a difficult time keeping her child with ASD on task. All trained peers/siblings
indicated that they would like to participate in peer/sibling-assisted activities in future.
When the aquatic program ended, a classroom teacher of a child with ASD approached the primary researcher and asked
for more information about the peer/sibling-assisted technique and asked for ways she could use in her classroom.
Additionally, the primary researcher received an unsolicited from a mother of a child with ASD noting improvements and a
developing friendship with his trained peer. Two mothers of children with ASD agreed that their children’s social
interactions increased with their participation in the study.

4. Discussion

This study demonstrated the effect of peer/sibling-assisted learning on the interaction behaviors and aquatic skills in
children with ASD in an after-school aquatic program. The predication that children with ASD in the experimental groups
would increase physical and social interactions with TD peers/siblings during peer/sibling-assisted condition was verified.
Trained peers/siblings assisted their paired children with ASD to produce improvements by increasing the physical and
social interactions during peer/sibling-assisted condition. An additional positive finding was that the occurrence of trained
peers and siblings appeared to influence all children with ASD socially during voluntary peer/sibling support condition. More
importantly, trained peers and siblings also took advantage of peer/sibling-assisted learning for enhancing their aquatic
skills.
Significant improvements on aquatic skills for trained peer/sibling-assisted children with ASD may be attributed to the
fact that the current study was essentially a teacher-directed program. The current aquatic program was 60 min in length
consisting 10 min of the warm-up period, 35 min of the main instruction (20 min and 15 min of the teacher-directed and the
peer/sibling-assisted conditions, respectively), and 15 min of games (voluntary peer/sibling support condition) at the end of
class. The trained peer/sibling was instructed on how to assist the child with ASD only during the 15 min of the peer/sibling-
assisted condition. Instructors may facilitate and monitor the peer/sibling-assisted condition from close by although they
never intervene directly with the children with ASD. The changes of social interaction behaviors were in fact that the children
interacted with one another. However, no evidence was found in this study to support the idea that compared to CG of
children with ASD, peer/sibling-assisted learning increases performance on aquatic skill acquisition for PG and SG of children
ASD. It can be assumed that activity engagement with assistance given by the trained peers/siblings during aquatic program
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223 1221

may have a positive impact on interaction behaviors, but not on motor skill outcomes yet. Furthermore, the instructors
provided aquatic skill instructions to both children with ASD and TD child to ensure that both children had adequately
learned the appropriate aquatic skills.
Effect of the present aquatic program on aquatic skills is similar to the finding reported in the previous research (Pan,
2010). Pan (2010) found that children with ASD showed improvement in percentage scores on four (II, III, IV, and V) and three
stages (II, IV, and V) of the HAAR checklist for groups A and B of children with ASD, respectively. The majority of children with
ASD increased their skills on introduction to water environment, rotations, balance and controlled movement, and
independent movement in water after 10 weeks training. The increment of scores for group A of children with ASD ranged
from 33.33% to 64.60%, and the increment of scores for group B of children with ASD ranged from 28.75% to 43.75%. In the
current study, all children with ASD showed a significantly larger improvement after 16-week training. The increment of
scores ranged from 30.00% to 51.79% for PG, from 28.57% to 57.14% for SG, and from 42.56% to 66.07% for CG. These
similarities perhaps due in part to the structured teaching model and curriculum-based assessment served as the basis for
the instructional process of the aforementioned research and the current study. Children with ASD learn in ways that are
unique to their TD peers. The current aquatic program identified each child’s unique strengths, weaknesses, and learning
style and then to use this information to build on strengths while modifying activities and creating accommodations for
weaknesses. The addition of extra structure for children with ASD such as visual schedules, clearer visual boundaries, and
more defined task cues helped them to be more successful in the program. Furthermore, individualized assessment allowed
the instructors to understand how best to present material and teach each child, and therefore, maximized potential positive
impacts of the current program.
The favorable effects of the peer/sibling-assisted learning on physical and social interaction behaviors are consistent with
previous studies in the educational literature, which have shown that peer- or sibling-mediated intervention increased
interaction behaviors in the area of PE (Klavina & Block, 2008) and social interaction behaviors in natural setting (Knott et al.,
2007; Taso & Odom, 2006). Activity engagement with assistance from teachers significantly decreased, and activity
engagement with trained peers or siblings significantly increased during peer/sibling-assisted condition for both PG and SG
of children with ASD. This suggests decreased dependence of the children with ASD on teachers and increased their
interactions with their peers or siblings. The data of social interactions demonstrated an accelerating trend across peer/
sibling-assisted and voluntary peer/sibling support conditions. This had a positive effect on social activity engagement for
both children with ASD and their TD peers/siblings, and supported the literature indicating that peer- or sibling-mediated
intervention is critical if they are to be placed in situations where they can observe, imitate, and receive discriminated
feedback and prompts (Harrower & Dunlap, 2001). These findings have shaped future research in inclusive physical activity
programs toward social outcomes in addition to motor or fitness measure in children with ASD.
Potential benefits of the current aquatic program for trained peers and siblings were also observed. Specially, all TD
children performed better on the stages II–V of the HAAR checklist after the 16-week training. The increment of scores
ranged from 31.43% to 73.81% for PG, from 27.14% to 61.90% for SG, and from 22.86% to 57.15% for CG. The positive influence
of peer/sibling-assisted learning on aquatic skills and social interactions parallels findings of Legrain, d’Arripe-Longueville,
and Gernigon (2003), which indicated that helping may have similar motor skill as well as social and behavioral advantages
for helpers in a sport setting. Together, these results indicate that peer/sibling-assisted program can contribute to a richer
environment than more conventional teaching approaches. This study suggests that elementary school-aged children
without a disability should be encouraged to take a more active approach to their learning and thus develop greater
academic engagement and successful collaboration as learners. Coaches and PE teachers could therefore pay attention to the
advantages of encouraging students to help each other with their learning rather than relying solely on the teacher.
Previous studies have demonstrated improvement in swimming and behavior skills as a result of a segregated swimming
program in children with ASD (Huettig & Darden-Melton, 2004; Pan, 2010; Yilmaz, Yanardag, Birkan, & Bumin, 2004).
Obvious differences in aquatic skills of children with ASD after the program were also seen in the current study. All of these
results suggest that aquatic program with structured supports, either in a self-contained or peer/sibling-assisted setting,
may be a fun alternative to land-based physical activity programs for improving social and motor skills in children with ASD.
The current study replicates previous research (Klavina & Block, 2008; Pan, 2010) and documents peer/sibling assisting as an
effective strategy for enhancing academic achievement (aquatic skills) and interaction behaviors. More important is the
demonstration on improving social interactions during voluntary peer/sibling support condition. The findings of this study
further suggest that instructors need to continue to address social goals for children with ASD in physical activity setting.
Benefits of using a peer/sibling-assisted learning program similar to the one used in this study are that it is easy to establish
and it benefits everyone involved. Instructors can share this information with parents and other professionals who want to
implement a similar peer/sibling-assisted program. The benefits for the peers/siblings as well as the children with ASD make
the use of peer/sibling assisted learning an even more desirable instructional strategy. With the current economic situation,
school districts often times are being forced to reduce support personnel. The use of trained peer/sibling or volunteers can be
a viable option for providing children with disabilities additional and individual support.
In conclusion, this quantitative analysis showed that participant children with ASD improved their interaction behaviors
and aquatic skills during an after-school inclusive aquatic program. Trained peers/siblings in aquatic settings could receive
some advantage from assisting others. Furthermore, this study addressed school- and community-based physical activity
programs, an opportunity that exists at all schools and communities and often times presents explicit challenges for children
with ASD and their TD peers and siblings to create physical, social, and behavioral interactions in and after school.
1222 C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223

Limitations of this study should be recognized. First, this study was limited in scope because the number of participants
was relatively small, and therefore were not necessarily representatives of the larger population of elementary school
children with ASD and their TD peers/siblings. Second, the repeated measurement design of all groups participated in all
three conditions over a 16-week period may confound the results as there are likely to be lasting effects that carry over from
one condition to another condition. Third, the presence of the researcher in the pool during each class may have elicited
reactivity by the participants. Finally, even though children with ASD were mild or high-functioning end of autism,
differences in social skills, behaviors, cognitive abilities, and motor skills were not evaluated and might have influenced
findings.
Future studies should examine the effect of trained peers/siblings on other components of the lesson, such as the fitness
component. Studies should also examine the relationship between interaction behaviors and fitness levels or physical
activity performance. Future studies should consider assessing the costs and benefits of implementing peer- or sibling-
assisted techniques to increase interaction behaviors and aquatic skills as well as other outcome measures. Additional
studies should examine children with different types and levels of disabilities and across a wider variety of settings and age
groups. It would also be valuable to examine systematically how the characteristics of the peers/siblings (e.g., age, gender)
are related to intervention effectiveness of the program. Finally, future research should use class wide peer tutoring so that
all students in the class have the opportunity to interact with students with disabilities.

Acknowledgments

This research was supported by NSC 97-2410-H-017-027-SSS and NSC 98-2410-H-017-025. The authors gratefully thank
Dr. Aija Klavina for helpful suggestions and for providing the CEPI-PE instrument and the complete definitions of codes with
examples. The authors also thank all children who participated in this study, parents of children for their supports, and
research assistants who helped with data collection and other contributions.

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders: Text revision (4th ed.). Washington, DC: American Psychiatric
Association.
Attwood, T., (1998). Asperger’s syndrome: A guide for parents and professionals. Philadelphia: Jessica Kingsley.
Blubaugh, N., & Kohlmann, J. (2006). TEACCH model and children with autism. Teaching Elementary Physical Education, 16–19.
Bowman-Perrott, L., (2009). Classwide peer tutoring: An effective strategy for students with emotional and behavioral disorders. Intervention in School and Clinic,
44, 259–267.
Cairney, J., Hay, J., Faught, B. E., Corna, L. M., & Flouris, A. D. (2006). Developmental coordination disorder, age, and play: A test of the divergence in activity-deficit
with age hypothesis. Adapted Physical Activity Quarterly, 23, 261–276.
Cantell, M. H., Crawford, S. G., & Doyle-Parker, P. K. (2008). Physical fitness and health indices in children, adolescents and adults with high or low motor
competence. Human Movement Science, 27, 344–362.
DePaepe, J. L., (1985). The influence of three least restrictive environments on the content motor-ALT and performance of moderately mentally retarded students.
Journal of Teaching in Physical Education, 5, 34–41.
Gilberts, G. H., Agran, M., Hughes, C., & Wehmeyer, M. (2001). The effects of peer delivered self-monitoring strategies on the participation of children with severe
disabilities in general education classroom. The Journal of the Association for Persons with Severe Handicaps, 26, 25–36.
Harrower, J., & Dunlap, G. (2001). Including children with autism in general education classrooms: A review of effective strategies. Behavior Modification, 25, 762–
784.
Houston-Wilson, C., Dunn, J. M., van der Mars, H., & McCubbin, J. (1997). The effect of peer tutors on motor performance in integrated physical education classes.
Adapted Physical Activity Quarterly, 14, 298–313.
Huettig, C., & Darden-Melton, B. (2004). Acquisition of aquatic skills by children with autism. Palaestra20(2) 20–25 and 45–46.
Humphries, K. M. (2008). Humphries’ assessment of aquatic readiness. Unpublished Master’s Thesis, Texas Woman’s University, Denton, Texas, USA.
Hutzler, Y., Chacham, A., Bergman, U., & Szeinberg, A. (1998). Effects of movement and swimming program on vital capacity and water orientation skills of
children with cerebral palsy. Developmental Child Neurology, 45, 367–370.
Klavina, A. (2007). The effect of peer tutoring on interaction behaviors in inclusive physical education. Unpublished Doctoral Dissertation, University of Virginia,
Charlottesville, Virginia, USA.
Klavina, A., & Block, M. E. (2008). The effect of peer tutoring on interaction behaviors in inclusive physical education. Adapted Physical Activity Quarterly, 25, 132–
158.
Klavina, A., & Selavo, L. (2006). Analyses of instructional, social and physical interaction in physical education (AISPI-PE). In J. Wittmanova (Ed.), In Proceedings:
European conference in adapted physical activity (pp. 49–).
Knott, F., Lewis, C., & Williams, T. (2007). Sibling interaction of children with autism: Development over 12 months. Journal of Autism and Developmental Disorders,
37, 1987–1995.
Legrain, P., d’Arripe-Longueville, F., & Gernigon, C. (2003). Peer tutoring in a sport setting: Are there any benefits for tutors? The Sport Psychologist, 17, 77–94.
Lieberman, L. J., Dunn, J. M., Mars, H., & McCubbin, J. (2000). Peer tutors’ on activity levels of deaf children in inclusive elementary physical education. Adapted
Physical Activity Quarterly, 17, 20–39.
Livesey, D., Mow, M. L., Toshack, T., & Zheng, Y. (2011). The relationship between motor performance and peer relations in 9- to 12-year-old children. Child: Care,
Health and Development, 37, 581–588.
MacKinnon, K., (2003). University of Central England Case Study: An evaluation of the benefits of Halliwick swimming on a child with mild spastic diplegia http://
www.halliwick.org.uk/downloads/casestudy01.PDF (retrieved 24.08.08).
Martin, H., (1981). The Halliwick Method. Physiotherapy, 67, 288–291.
Mcduffie, K. A., Mastropieri, M. A., & Scruggs, T. E. (2009). Differential effects of peer tutoring in co-taught and non-co-taught classes: Results for content learning
and student–teacher interactions. Exceptional Children, 75, 493–510.
Mesibov, G., Shea, V., & Schopler, E. (2004). The TEACCH approach to autism spectrum disorders. New York: Springer.
Pan, C. Y., (2010). Effects of water exercise swimming program on aquatic skills and social behaviors in children with autism spectrum disorders. Autism:
International Journal of Research and Practice, 14, 9–28.
Pan, C. Y., Tsai, C. L., & Chu, C. H. (2009). Fundamental movement skills in children diagnosed with autism spectrum disorders and attention deficit hyperactivity
disorder. Journal of Autism and Developmental Disorders, 39, 1694–1705.
C.-H. Chu, C.-Y. Pan / Research in Autism Spectrum Disorders 6 (2012) 1211–1223 1223

Piek, J. P., Baynam, G. B., & Barrett, N. C. (2006). The relationship between fine and gross motor ability, self-perceptions and self-worth in children and adolescent.
Human Movement Science, 25, 65–75.
Provost, B., Heimerl, S., & Lopez, B. R. (2007). Levels of gross and fine motor development in young children with autism spectrum disorder. Physical and
Occupational Therapy in Pediatrics, 27, 21–36.
Skinner, R. A., & Piek, J. P. (2001). Psychosocial implications of poor motor coordination in children and adolescents. Human Movement Science, 20, 73–94.
Staples, K. L., & Reid, G. (2010). Fundamental movement skills and autism spectrum disorders. Journal of Autism and Developmental Disorders, 40, 209–217.
Strickland, J., Temple, V. A., & Walkley, J. W. (2005). Peer tutoring as an instructional methodology to improve fundamental movement skills. The ACHPER Healthy
Lifestyles Journal, 52(2), 22–26.
Taiwan Executive Yuan Council of Labor Affairs. (2010). Levels of disabilities http://opendoor.evta.gov.tw/sub.aspx?a=0004437&p=0000107 (retrieved 31.01.10).
Taiwan Executive Yuan Department of Health. (2010). Hospitals for the identification of disabilities http://www.doh.gov.tw/CHT2006/DM/DM2_p01.aspx?class_-
no=211&now_fod_list_no=6753&level_no=3&doc_no=41058 (retrieved 31.01.10).
Taso, L. L., & Odom, S. L. (2006). Sibling-mediated social interaction intervention for young children with autism. Topics in Early Childhood Special Education, 26,
106–123.
Ward, P., & Ayvazo, S. (2006). Classwide peer tutoring in physical education: Assessing its effect with kindergartners with autism. Adapted Physical Activity
Quarterly, 23, 233–244.
Yilmaz, I., Yanardag, M., Birkan, B., & Bumin, G. (2004). Effects of swimming training on physical fitness and water orientation in autism. Pediatrics International, 46,
624–626.

You might also like