EDUCATION IN A COMPETITIVE AND GLOBALIZING WORLD SERIES










SPECIAL EDUCATION IN THE
21ST CENTURY








MARYANN T. BURTON
EDITOR















Nova Science Publishers, Inc.
New York

Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.

Copyright © 2010 by Nova Science Publishers, Inc.


All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.

For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com

NOTICE TO THE READER
The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or
implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of
information contained in this book. The Publisher shall not be liable for any special,
consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or
reliance upon, this material.

Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise
contained in this publication.

This publication is designed to provide accurate and authoritative information with regard to the
subject matter cover herein. It is sold with the clear understanding that the Publisher is not
engaged in rendering legal or any other professional services. If legal, medical or any other expert
assistance is required, the services of a competent person should be sought. FROM A
DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE
AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.

Library of Congress Cataloging-in-Publication Data

Special education in the 21st century / [edited by] MaryAnn T. Burton.
p. cm.
Includes index.
ISBN 978-1-61728-489-2 (E-Book)
1. Special education. I. Burton, MaryAnn T.
LC3965.S646 2009
371.9--dc22
2009024621



Published by Nova Science Publishers, Inc. New York
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.









CONTENTS

Preface ix
Chapter 1 Special Education and Early Arithmetic Knowledge-Building 1
Göta Eriksson
Chapter 2 Modelled, Free Play and Toy Type: Associations with Sociability,
Play and Language Usage Among Intellectually Disabled and
Typically Developing Children 37
Vickii B. Jenvey and Heidi L. Jenvey
Chapter 3 The Necessity of Measures of Procedural Integrity to Ensure
Effective Delivery of Evidence-Based Interventions for
Children with Autism Spectrum Disorders 67
Thomas P. Kitchen, Phillip J. Belfiore and Holly L. Kitchen
Chapter 4 Comprehensive Intervention for High-Functioning Autism
Spectrum Disorders: An In-Depth Case Study 95
Jennifer A. Toomey, Christopher Lopata, Martin A. Volker and
Marcus L. Thomeer
Chapter 5 Inclusion: Theory and Practice for Ex-Students of Special Schools 119
Garry Hornby
Chapter 6 The Contribution of Temperament and Intellectual Functioning
to Social Behaviour in Children 137
Efrat Zion and Vickii B. Jenvey
Chapter 7 Ideas for Teachers Working with Children who Have Medical
Conditions 155
L. A. Nabors, B. T. Olsen and E. Henderson
Chapter 8 Transition for High School Special Education Students to Post-
Secondary Environments: Best Practices, Benefits, and Barriers 169
Jennifer L. Koehler, Tanya L. Eckert, Adrea J. Truckenmiller,
Jennifer L. Rymanowski and Elizabeth A. Koenig


Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Contents viii
Chapter 9 Special Education for Children with Autism Spectrum Disorder
(ASD): Implementation of Programmes Based on Innovative
Technology 183
A. Battocchi, G. Esposito, S. DeFalco and P. Venuti
Chapter 10 Trained Peers' Interactions With Children With Autism During
Play 195
Serene Hyun-Jin Choi and Timo A. Nieminen
Index 225


Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
In: Special Education in the 21st Century
Editor: MaryAnn T. Burton, pp. 195-224
ISBN 978-1-60741-556-5
© 2010 Nova Science Publishers, Inc.
Chapter 10
TRAINED PEERS’ INTERACTIONS WITH CHILDREN
WITH AUTISM DURING PLAY
Serene Hyun-Jin Choi∗and Timo A. Nieminen†
School of Physical Sciences, The University of Queensland,
Brisbane, QLD 4072, Australia
Abstract
In view of the trend towards increased inclusion, the placement of children with
autism into inclusive educational settings is likely to increase. Such placement typ-
ically fails to result in social interaction with typically developing peers, which is
usually explained as a result of deficits of social skills in the children with autism.
However, social interaction is a reciprocal process, and it is reasonable to expect that
the difficulty at least partly lies with the other child.
Accordingly, a training program, to teach typically developing children how to
interact with children with autism in indoor and outdoor play activities was devel-
oped. The effectiveness of the training program was tested experimentally (a modified
subjects-as-their-own-controls design) in dyad groups, with the same children with
autism each playing with a trained peer and an untrained peer in dyads. This mod-
els a peer-mediated play intervention, and the difference in success between trained
and untrained dyads allows the effectiveness of the peer training to be determined.
The level of social interaction displayed by the children with autism was measured
and compared in the dyads. As the cognitive aspect of play, using Piaget–Smilansky
types of categories, is a commonly-used measure of effect, this was also measured and
compared in the dyads.
It was demonstrated clearly and conclusively that when typically developing peers
were systematically trained, they were more effectively and consistently able to play
and interact with the children with autism. Untrained peers, on the other hand, were
not able to interact effectively with the children with autism, and the level of social in-
teraction did not improve over time, indicating that proximity alone is ineffective. This
demonstrates that the peer training program was both effective, resulting in statistically
improved social interaction, and necessary, as no improvement occurred without train-
ing. However, the measurement of the cognitive play did not provide clear answers,
∗E-mail address: serene@aanet.com.au
†E-mail address: timo@physics.uq.edu.au
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
196 Serene Hyun-Jin Choi and Timo A. Nieminen
implying that the Piaget–Smilansky cognitive play categories are a poor indicator of
success of intervention programs for children with autism.
1. Introduction
In view of the trend towards increased inclusion, the placement of children with autism
into inclusive educational settings is likely to increase. An important benefit of such in-
clusive placement is the opportunity for the children with autism to socially interact with
typically developing peers. This is an important issue, since a child can influence the social
development of other children, either positively or negatively. Positive and supportive inter-
action with peers contributes to development in both children with and without disabilities,
as discussed by Hartup (1999), Piaget (1959), and Vygotsky (1962). On the other hand,
peers can negatively influence children by reinforcing inappropriate behaviour, for exam-
ple by paying attention to aggressive disruptive behaviours in children with developmental
disorders (Marcus, Vollmer, Swanson, Roane, & Ringdahl, 2001). Therefore, the sense of
well-being, and quality of life, of students with disabilities at educational settings might be
more dependent on supportive peer relationship in terms of inclusive education programs,
rather than either integrated or segregated placement issues (Allodi, 2000).
However, when children with autism are placed in inclusive settings, social interaction
typically fails to result—physical proximity is insufficient. This is not so surprising, since
difficulty in social interaction has been recognised as a major diagnostic sign of autism;
manifesting as interaction with typically developing peers being either negative or rare.
Accordingly, much research has focussed on improving social interaction between children
with autism and peers and others; this field of research has been reviewed by McConnell
(2002) and Choi (2007). Not unexpectedly, no panacaea has emerged. While there has
been some success, further progress towards training methods or interventions that can en-
hance social interaction while minimising participation by skilled personel and other adults
remains highly desirable.
Theories and hypotheses on this difficulty in interaction have focused on the difficulty
caused by children with autism, with the lack of social interaction usually explained as a
result of deficits of social skills in the children with autism. However, social interaction
is a reciprocal process, and it is reasonable to expect that the difficulty that children with
autism experience in social interaction with typically developing children is partly due to
the typically developing children being unable to recognize or understand social cues and
responses used by the children with autism.
Therefore, a training program to teach typically developing children how to interact
with children with autism was developed (Choi, 2005). The training program was designed
so that it would be possible for a regular classroom teacher to train the typically developing
children in the class in preparation for the inclusive placement of a child with autism. The
experimental test of this training program was the primary focus of the research described
here.
A modified subjects-as-their-own-controls design was used, with the same children with
autism each playing with a trained peer and an untrained peer in dyads, in a naturalistic set-
ting. This simplified model of peer interaction during non-class school time allowed reliable
quantitative results to be obtained. It should also be noted that this test also indicated the
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 197
expected performance of the training program if used in a peer-mediated play intervention.
The training program was clearly shown to be effective. The design of the training
program and its experimental test, and the results of that test, are described below.
2. Method
2.1. Measurement of success
Since the effectiveness of the peer training was determined by measurement of the differ-
ence in success between trained and untrained dyads, it will be beneficial to first consider
the methods by which the level of success of such an intervention can be measured, before
considering the details of the experimental design.
Perhaps the most common method for quantitative measurement of success of peer in-
terventions is measure the “level” of cognitive behaviour by classification using the Piaget–
Smilansky categories (Piaget, 1962; Smilansky, 1968)—since the pre-eminent theory relat-
ing play and cognitive development is that of Piaget (1952, 1977), it is hardly surprising
that taxonomies and classification schemes, which provide a bridge between the concepts
embodied in the theory and the practical matters of measurement and experiment, based on
his theory and its derivatives are in common use. However, it must be noted that Piaget
deliberately ignored social factors. As Piaget (1962, p. 68) explained, “We are of course
entirely in agreement that thought cannot be explained without recourse to social factors,
but the general concept of ‘social life’ seems to us inadmissible in psychology. ‘Society’ is
neither a thing nor a cause, but a system of relationships, and it is for the psychologist to
classify these relationships and analyse separately their respective effects”. Hetherington,
Parke, and Locke (1999) pointed out that the focus of Piagetian theory is on the set of
innate cognitive abilities or limitations of children associated with each stage and as such
social and cultural influences on children’s cognitive development are largely not addressed.
Thus, one should consider carefully the relevance of a Piagetian classification for a mea-
surement of success of a social intervention. On the other hand, such a classification allows
comparison of results with studies making use of similar schemes.
Therefore, the cognitive level of play was measured using the Piaget–Smilansky cat-
egories, but it was also decided to make a more direct measurement of the level of so-
cial interaction. Unfortunately, while the Piaget–Smilansky categories are widely-used and
widely-accepted, amounting to a de facto standard set of categories, there is no equivalent
measure of social interaction behaviour (especially for children with autism). The six cat-
egories (unoccupied play, onlooker play, solitary play, parallel play, associative play, and
cooperative play) established by Parten (1932) are widely used for the measurement of the
social aspect of peer play in typically developing children (Farran & Son-Yarbrough, 2001;
Fox, 1996; Johnson, Christie, & Yawkey, 1999; Pugmire-Stoy, 1992; Saracho, 1993), and
have even been used in some studies of children with developmental delay (e.g., Brophy
& Zukowski, 1984) or with autistic spectrum disorder (e.g., Anderson, Moore, Godfrey, &
Fletcher-Flinn, 2004; Jahr, Eikeseth, Eldevik, & Aase, 2007; Yang, Wolfberg, Wu, & Hwu,
2003), and Parten’s categories appear to be the closest that there is to a standard set of cat-
egories for the classification for peer play behaviour. However, Parten’s categories do not
seem to be sufficient to classify meaningful types or levels of social interaction of children
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
198 Serene Hyun-Jin Choi and Timo A. Nieminen
with or without disabilities—Parten’s categories do not allow the measurement of social in-
teraction occurring without play. Even in play-centred settings, this is a serious limitation;
while social interaction during play sessions will usually be part of play activities, non-play
social interaction can occur, and must be measured. If play is not the primary activity of the
children whose behaviour is being measured, this limitation makes Parten’s categories fun-
damentally unsuitable. This has been recently noted by Anderson et al. (2004) with some
surprise, since the categories are in wide use.
Especially since the cognitive level of play would already be measured using the Piaget–
Smilansky cognitive categories, it was decided to design new categories to classify social
interaction and play with peers in children with autism to allow for the hesitant, vague,
uncertain, and brief social behaviours that can be expected. The new social interaction
categories were to be independent of the actual play, so as to provide an independent (of the
cognitive level of play) dimension of measurement.
This development of a taxonomy of social interaction can be carried out in conjunction
with the development of a binary key as a practical tool for the classification of the observed
behaviour, as has been described in detail elsewhere (Choi, 2005; Nieminen & Choi, 2008).
The key for the classification of the cognitive level of play is shown in figures 1 and 2,
and the key for the classification of the level of social interaction is shown in figures 3 and
4. The keys are shown in both the more traditional typeset “list” form (figures 1 and 3)
and the easier-to-use graphical tree form (figures 2 and 4). The keys themselves provide a
compact definition of the categories; for example, the active–low category can be specified
as: the child with autism shows non-negative social responses or initiation with the peer,
with at least some initiation by the child with autism; however, while the the interaction
is reciprocal, the behaviour of the child with autism is unclear; and the peer shows a non-
negative response to initiations by the child with autism.
Of these measurements of the levels of cognitive play and social interaction displayed
by the the children with autism, the level of social interaction provides a direct measurement
of the immediate success of the peer-mediated play intervention, and is the best available
indicator of possible long-term benefit for the children with autism.
2.2. Design
In principle, it is quite straightforward to measure the effectiveness of a training program
for typically developing children to improve their interaction with children with autism:
simply compare the levels of social interaction shown by children with autism when playing
with trained and untrained typically developing peers. In practice, however, inter-subject
variability would require a large number of subjects before a statistically significant result
could be established.
Therefore, a standard procedure is to make use of quasi-experimental methods, such as
using a baseline/treatment design (AB design), where the subjects before treatment (in this
case, before the training program) act as a control group. Since both the experimental group
and the control group are made up of the same subjects, the effect of intersubject variabil-
ity is virtually eliminated. This was the method adopted here, with typically developing
children acting first as untrained peer play partners, undergoing training, and then acting as
trained peer play partners.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 199
1. Is the child obviously playing?
(a) Yes—go to 4
(b) No—go to 2
2. Is the child exploring or examining play objects or the play partner’s play?
(a) Yes—Exploratory
(b) No—go to 3
3. Is the behaviour typical stereotype behaviour?
(a) Yes—Stereotype behaviour
(b) No—Non-play
4. Does the play consist only of play objects being used in a physical manner?
(a) Yes—go to 5
(b) No—go to 6
5. Are the play objects being organised spatially?
(a) Yes—Constructive play
(b) No—Functional play
6. Are rules obviously being followed?
(a) Yes—Rule-governed play
(b) No—Symbolic play
Figure 1. Key for classification of cognitive play (list format)
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
200 Serene Hyun-Jin Choi and Timo A. Nieminen
Figure 2. Key for classification of cognitive play (tree format)
However, and this is especially the case when testing the effect of a training program,
any change could be due to the passage of time—improvement through practice, rather than
through training. Since the trained peers cannot be untrained, it was not possible to use a
baseline/treatment/reversal design (ABA design) to overcome this difficulty. It was then
necessary to use a second group of peers who remained untrained to measure the amount of
improvement that resulted from practice. In order to minimise the effect of variation among
the children with autism, these untrained peers played with the same children with autism
as the trained peers.
This left open the possibility that improvement in social interaction with the untrained
peers could have been due to learning by the children with autism, from the trained peers.
Although it was realised that a strong crossover effect of this type, if it occurred, would
make it impossible to achieve the primary goal—measurement of the effectiveness of the
training program—such rapid learning of social skills by the children with autism was con-
sidered highly unlikely, and in any case, if it did occur would be an observation of such
importance as to compensate for the failure to determine the effectiveness of the training
program.
In order to be able to use statistical methods to determine the likelihood that any change
seen due to the training was genuine, the social interaction and play behaviour of the chil-
dren was autism, in play sessions with trained and untrained peers, was recorded over mul-
tiple sessions both before and after the training, so that sufficient data were available to
allow accurate determination of the mean levels of social interaction and cognitive play
behaviour.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 201
1. Does the child with autism show non-negative response or initiation?
(a) Yes—go to 3
(b) No—go to 2
2. Does the child with autism show negative social behaviour?
(a) Yes—Negative interaction
(b) No—No interaction
3. Is the child with autism only responding to the peer (i.e. not initiating)?
(a) Yes–go to 6
(b) No—go to 4
4. Is there any non-negative response from the peer?
(a) Yes—go to 5
(b) No—Unilateral
5. Is there clear and reciprocal interaction?
(a) Yes—Active–high
(b) No—Active–low
6. Is there clear initiation by the peer and clear response by the child with autism?
(a) Yes—Passive–high
(b) No—Passive–low
Figure 3. Key for classification of social interaction (list format)
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
202 Serene Hyun-Jin Choi and Timo A. Nieminen
Figure 4. Key for classification of social interaction play (tree format)
2.3. Participants
After ethical clearance and consent by involved paerties, a total of nineteen children—five
children with autism and fourteen typically developing peers—were recruited.
The five children diagnosed with autism were all boys, aged from three years and one
month to six years and nine months at the time of their participation. These children autism
were recruited from two Special Education Developmental Units (SEDU-A and SEDU-B)
and one Special Education School (SES) associated with SEDU-A. They were all Cau-
casian and from families where English was the only language spoken at home. All the
five children with autism had been diagnosed by a local psychologist. Four of them (CA1,
CA2, CA4, and CA5) attended their SEDU on a part-time basis for special educational
services with other children with developmental delay. The other boy with autism (CA3)
was enrolled in the SES on a full-time basis with other students with special needs. A brief
summary of the participating children with autism is give in table 1
Table 1. Summary of participants: Children with autism
Designation Gender Age
CA1 male 3 years, 1 month
CA2 male 5 years, 5 months
CA3 male 6 years, 9 months
CA4 male 4 years, 8 months
CA5 male 5 years, 9 months
Fourteen typically developing peers (seven boys and seven girls, aged eight to nine
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 203
years, designated P1 through to P14, with boys being the odd-numbered peers, and girls
the even-numbered peers) were recruited from three grade three (year three) classes in two
state primary schools (SPS-A and SPS-B). All of the children who acted as typically devel-
oping peers volunteered to take part in the study, following initial selection by their teachers
(Jackson & Campbell, 2008). Apart from one girl (P8) who was non-Caucasion of unknown
background (possibly Maori or other Polynesian), all other peers were Caucasian. One boy
(P7) had mild paraplegia.
Of these fourteen cross-aged peer players, ten acted as regular peer players (two for
each child with autism). The other four students (P6, P7, P12, and P13) were substitute
or reserve peers, to replace regular peers who were not present for a scheduled session or
withdrew their participation. In the event, P13 did not participate in any of the play sessions.
Therefore, no details on P13 are reported in this study. Since the other reserve peers (P6,
P7, and P12) did have contact with the children with autism, their profile and screening
results are reported here. A brief summary of the typically developing peers is given in
table 2 (excluding P13).
Table 2. Summary of participants: Typically developing peers
Designation Gender Age
P1 male 7 years, 11 months
P2 female 8 years, 9 months
P3 male 8 years, 8 months
P4 female 7 years, 10 months
P5 male 7 years, 6 months
P6 female 8 years
P7 male 7 years, 8 months
P8 female 8 years, 2 months
P9 male 7 years, 11 months
P10 female 7 years, 11 months
P11 male 8 years, 4 months
P12 female 7 years, 10 months
P14 female 8 years, 4 months
All schools and special education settings had similar Index of Relative Socio-economic
Disadvantage (IRSD) scores (Pink, 2008) from 952 to 1043, ranging from the 25% quan-
tile to the 75% quantile, which places them in “medium disadvantage” areas. The socio-
economic status of students in these two SPS was identified by school staff as mostly work-
ing class.
2.4. Settings and materials
Play and interaction in both indoor and outdoor settings were measured in this study. These
settings were simulated naturalistic settings, providing a high degree of both control and
generalizability to real-life situations. Both indoor and outdoor settings were used since
children in inclusive education could be expected to play together in both indoor and out-
door settings on a daily basis. Most similar studies were restricted to indoor settings only,
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
204 Serene Hyun-Jin Choi and Timo A. Nieminen
presumably on the grounds of experimental practicality. Thus, it was possible that the out-
door settings might provide especially new and interesting results.
The children with autism and their peer partners came from two Special Education De-
velopmental Unit (called SEDU-A and SEDU-B) and State Primary School (called SPS-A
and SPS-B) pairs. For the SEDU-A–SPS-A pair (dyads with CA1, CA2, and CA3), the play
sessions were held at SPS-A. Teacher aides in SEDU-A escorted the children with autism to
SPS-A. For the SEDU-B–SPS-B pairs (dyads with CA4 and CA5), the play sessions were
held at SEDU-B. The researcher of this study escorted the typically developing peers to
SEDU-B. These choices were necessitated by the practical availability of space.
In SPS-A, the indoor play setting was a space of approximately four and a half by five
metres which was partitioned off by desks and chairs from a larger function room. In the
SEDU-B, the indoor play setting was a room of approximately four metres by four metres
in size.
The outdoor play settings at SEDU-B and SPS-A were similar in that both included a
sandpit and an adequately-sized playground. The outdoor play setting at SEDU-B also had
a play area used for playing with outdoor toys, distinct from the playground. The SEDU-B
playground was also smaller, approximately half the size (but still of adequate size), and
access was dependent on usage by other classes.
For the indoor play sessions, four sets of play materials were chosen from those avail-
able at the SEDU attended by the children with autism and placed on the floor. The play
materials were selected to focus on the four cognitive play categories of functional, con-
structive, symbolic, and rule-governed play. Each set of toys was considered as a repre-
sentation for each play category. For example, a cars and road set was for functional play,
blocks for constructive play, a kitchen set for symbolic play, and matching-cards for rule-
governed games (see table 3).
For outdoor play sessions, some sandpit play materials were chosen from the two SEDU
(see table 4). In addition, playground facilities such as slides (and monkey bars at SPA-A)
were used. Other outdoor play equipment such as a baseball set and a tenpin bowling set
were set up in SEDU-B (see table 5).
Table 3. Indoor play materials
SPS-A (CA1–3) SEDU-B (CA4–5)
Cars and road Cars and road
Doll and bassinet Doll and clothes
Doctor set Doctor set
Ring-toss games Lego train and track
Plastic chains and connectors Dice
Skittles and ball Card game set
Plastic farm animals
Cooking set
Drinking straws
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 205
Table 4. Sandpit play materials
SPS-A (CA1–3) SEDU-B (CA4–5)
Sieves Sieves
Shovels Shovels
Containers Containers
Trucks Trucks
Sand-turned wheels Stove, frypan, and spoons
Table 5. Playground play materials and facilities
SPS-A (CA1–3) SEDU-B (CA4–5)
Monkey bars T-ball
Climbing bar Hanging ball and rackets
Climbing net Skittles and ball
Tyre tunnel Hoops
Baseball set
2.5. Play training for peers
Half of the typically developing peers were chosen randomly to receive a series of play
training sessions (Trained Peers—TP), the other half, the comparison group, were not
trained (Untrained Peers—UP). In this peer training phase, the play training sessions for the
randomly chosen peers were arranged according to their school schedule, and implemented
in the same places for indoor and outdoor sessions where they played in the pre-training
phase. Untrained peer players kept their regular lessons in their classrooms.
The peer training was implemented in group sessions. However, if a TP missed a play
training session due to sickness or some other reason, an individual session was arranged
for him or her. A total of 12 peer training sessions was run at each SPS and each play
training session lasted approximately 30 to 40 minutes.
The training program focussed on disability awareness and play interaction skills that
were modified from peer approach interventions (see table 6).
Table 6. Training of peers
Structure Contents
Being friends Characteristics of friends and the importance of friend-
ship
Play interaction skills General play activities and behaviours
Different people General and/or specific differences in people
Play with different friends Verbal and non-verbal social communication skills
Understanding autism Comparing general behaviours between children with-
out autism and with autism
Play practices Elaborating the play interaction skills to play with play
partner with autism in indoor and outdoor settings
Total twelve specific play interaction skills were taught to the trained peers. Eleven play
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
206 Serene Hyun-Jin Choi and Timo A. Nieminen
interaction skills were adopted from Pivotal Response Training (PRT) (Pierce & Schreib-
man, 1995) as follows: getting attention from the child with autism by using eye-contact,
touching, and speaking (“Hello, [name of the child]”), allowing the children with autism to
choose play materials (e.g., “What do you want to play with, this car or this doll?”), paying
attention and waiting (“Please, show me how to use this car”), demonstrating play activi-
ties, including verbal statements (e.g., “My car goes on the road, just like this, brrrrum!”),
suggesting activites to children with autism (“Let’s play with the car”), turn-taking (“It’s
my turn” or “It’s your turn”), narration of play activity to children with autism (e.g., “This
is very spicy pizza,” or “You drive a car to go to the shop.”), providing help to children
with autism (“Push it more. I can help you.”), sharing activity (“We can use the ball to-
gether”), explaining (“Stand up in the circle, then roll it in the circle”), and reinforcement
of attempts to encourage and extend appropriate play behaviours and social skills in chil-
dren with autism (e.g., “Well done!”). One additional play interaction skill, asking for help
(“Can you help me? Push the block more”), was also included, with the intent of promot-
ing cooperation skills in the children with autism by sharing activities (Liebal, Colombi,
Rogers, Warneken, & Tomasello, 2008).
The training consisted of watching videotapes, discussions, specific instruction, mod-
elling, role-playing, feedback, worksheets, self-monitoring and feedback ( ‘Play Note’), and
rewards (see table 7). Videotapes that show general information about friendship, differ-
ences in people, and typical behaviours of children with autism were used to help the peers’
understanding of play interaction with children with autism. Relevant segments of educa-
tional films were prepared to fit within the limited time of play training session. Discussion
about play and social interaction was guided in a child player-centred manner. The dis-
cussion was supported and facilitated by simple questions focussed on the theme of each
session. The topic of discussion was started with general friendship in typically developing
children’s lives, and expanded to special friendships with children with disabilities, espe-
cially children with autism. Also, it focused not only on the notion that what children with
disabilities are not able to do due to their disabilities, but what children with disabilities are
able to do with their disabilities.
Table 7. Teaching strategies in peer play program
Methods Aims
Audio-visual materials To understand related topic and to monitor own play
Discussion To clarify the topic
Specific instruction To know interaction behaviours in children with autism
Modelling To demonstrate appropriate or inappropriate play be-
haviours
Role play To realize abilities and different needs of children with
disabilities and to develop their own play interaction
skills
Feedback To evaluate own performance
Worksheet To summarize and clarify what has been learned
Play Note Self-monitoring method by scoring play activities
Rewards To maintain and facilitate motivation
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 207
In the peer training, peers were guided to develop ways to play and interact with children
with autism. The peers exchanged their experiences and ideas of play with their siblings or
other friends in order to use these when playing with the children with autism. Role-playing
was also used to assist the peer play partners to understand more about playing situations.
In these role-play sessions, peers were ‘given’ a disability (such as being blindfolded to
emulate visual impairment) so that they could more easily empathise with the children with
disabilities. While peers experienced these disabilities, they could recognize the abilities
and different needs of individuals with disabilities from individuals without disabilities.
The worksheet enabled peers to summarize the session.
On the general understanding and experiences of dealing with people with disabilities,
peers practised and elaborated the play interaction skills by performing peer-adult, peer-
doll, and peer-peer in role-playing. When one peer demonstrated the play interaction skills,
other peers monitored him or her and allocated them a score on the Play Note. This moni-
toring activity was also carried out visa versa. When one peer demonstrated play interaction
skills, others identified what skill was used and their score on the Play Note. Eventually,
each peer performed the play interaction skills with an accuracy level aove 80% to play and
interact with an adult who took the role of a child with autism. All these practices were
carried out in a playful mood.
After the peer training phase, five more training sessions were performed in the post-
training phase. These on-going training sessions were given to the trained peers (TP) on
a fading schedule. It commenced with two days play framework, with a training session
following the third day’s play day. Each consecutive playtime frame had an additional day
added before the ‘on-going training’ day. These on-going training sessions lasted for about
20 minutes during lunch or recess time.
In the on-going training, the twelve play interaction skills were written on a piece of
paper and read by the TP, before they played with the children with autism. After the play
session, the TP watched videotapes that showed their play session with the children with
autism, and monitored their play interaction skills on the Play Note. If they missed the on-
going training session due to sickness or otherwise, an individual session was scheduled.
Also, untrained peers watched their own play sessions of tapes. Untrained peers were not
provided with the play interaction skills on the Play Note. All TP and UP watched their
‘play movies’ on the videotapes.
2.6. Procedure
After an initial profiling and screening phase, indoor and outdoor play sessions, covering a
pre-training phase, a peer training phase, and a post-training phase) were held two days per
week for each dyad, for approximately six months. Unfortunately, CA4’s regular trained
peer (TP10) withdrew after the peer training phase (and continued to participate in this
study as a reserve trained peer). TP10 was replaced by TP12 as the regular trained peer
player for CA4 during post-training phase. The data from these unmatched dyads (CA4–
TP10 and CA4–TP12) were collected. Even though the data from these unmatched dyads
could not provide conclusive evidence of improvement due to training (since improvement
from CA4–TP10 before training to CA4–TP12 after training could be due to TP12 being
more able to interact with CA4, independent of the training), it would still be useful to
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
208 Serene Hyun-Jin Choi and Timo A. Nieminen
check whether or not these data support the conclusions drawn from the matched dyads.
While the room in the SEDU-B was always available for indoor play sessions of this
study, two other rooms had to be used occasionally due to the unavailability of the regular
room and space in the SPS-A. Because this study was implemented in naturalistic education
settings for about seven months, availability of the room and space for this study depended
on the educational contexts such as extra room facilities for other educational programs and
services, school time schedule, and school activities, requiring occasional changes of room.
Such changes affected both trained and untrained peers equally.
While playground facilities in SPS-A were included, playground apparatus in SEDU-
B was not used, for two reasons. Firstly, some apparatus, such as a swing bar, were too
low for the typically developing peers. Secondly, this avoided conflict with use of those
facilities by other children at SEDU-B, who used them on their regular schedule. Instead,
other playground play materials, such as a T-ball, hoops, and a baseball set, were set up
near the sandpit area for outdoor play sessions at SEDU-B.
Each child with autism was alternatively paired with an untrained peer (untrained dyad)
and trained peer (trained dyad). Counterbalancing was generally maintained to minimize
the effect of the problem of same order implementation through alternating sequence of
untrained and trained dyad as well as alternating sequence of indoor and outdoor setting.
However, sometimes, one of the players was not available. Then, it was necessary to change
the sequence again.
At the start of a play session, each dyad was told “It is play time together! You can
play with these toys and have fun!” Each dyad had two play sessions, one indoor and
one outdoor, on each day they played. Therefore, each child with autism participated in
four play sessions, two with the trained peer, and two with the untrained peer, on each day
they played (twice per week). Each session lasted for about six minutes—six minutes was
the standard play session length, but variations did occur, as discussed below. Where the
session exceeded six minutes in duration, data were only recorded for the first six minutes.
In the play sessions, the dyad was reminded once or twice about unused available play
materials if they used only one item for more than half the time of the session (3 minutes).
Occasionally, other teaching materials were nearby, having been prepared for other children
by teaching staff, in which case, near the play settings. Then, the dyad was reminded which
play materials were available in the play session.
When one child in the dyad did not want to keep playing in the session, the child was
encouraged to continuing playing. After one or two minutes, if the child still did not want
to keep playing, the session was stopped. On the other hand, if both children in the dyad
group wanted to keep playing longer than six minutes, they were told that they would have
more opportunities to play together on other days. If they still insisted on playing longer,
they were allowed to play for about one or two minutes more. However, this extra play time
was only requested a couple of times from the trained dyad of CA4 in the late post-training
play phase.
Given that children with autism are especially vulnerable to disruption of their routines
(American Psychiatric Association, 2000), if a regular peer was unable to attend the sched-
uled play session, one of the reserve trained or untrained peers was used as a substitute.
This was important to minimise disruption for the children with autism. No notable disrup-
tion was observed as a consequence of the substitute peer play partner. Play sessions with
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 209
substitute peers were not counted as play sessions for the regular dyad in data analysis,
since the dyad was different. The play sessions with substitute peer were still all recorded,
even though the data were not to be used in this study.
2.7. Data collection and analysis
Indoor and outdoor play sessions of each dyad were recorded using a compact video cam-
era, and then transferred to VHS videotapes and coding was performed using an editing
machine which provided high-quality stills and accurate rewinding and fast-forwarding.
The dyad data on videotapes were classified into the two sets of behaviour categories de-
scribed earlier: one measuring the cognitive play, and the other measuring the level of social
interaction displayed by the child with autism as a function of time for each session.
The time at which the cognitive or social behaviour changed from one category to an-
other was recorded to the nearest second, giving a real-time coding of behaviour. Classifi-
cation of the behaviour from the videotapes was performed in two separate passes through
the tapes: once for the cognitive behaviour, and again for the social behaviour. Videotapes
were observed in a randomised order. The graphical tree versions of the cognitive and social
keys were used for the classification. After all videotapes had been coded, some randomly
selected videotapes were re-coded. The level of agreement between the original coding and
the re-coding was 97%. The major part of the small amount of disagreement was due to
small differences in the times at which one behaviour changed to another. It was also found
that stretching or expansion due to heating in the VCR of the videotapes may have been
responsible for a large part of these time differences. Accounting for this, the coding–re-
coding agreement could have been better than 98%. All coding was performed by a single
coder.
The total durations occupied by each category of behaviour of the child with autism
in matching dyads are of greatest interest; these were calculated for each session, and are
hereafter given as proportions of the total session. The data that is analysed here consists
of the proportion of the total time for each six minute (360 second) play session occupied
by each cognitive and social category. That is, each category has a duration of between 0
and 1, where 0 indicates a complete absence of the behaviour, and 1 would indicate that
it was the only behaviour displayed during that entire session. For example, a duration of
0.4 would mean that a total of 0.4 × 360 = 144 seconds was occupied by that behaviour
(assuming that the session was the standard 360 seconds long). Note that the sum of the
proportions of all of the cognitive or social categories for a single must be equal to 1, and
an increase in one category must be accompanied by decreases in other categories.
For a simplified overall analysis, it is very useful to condense the categorical data into
a single overall cognitive score and social score. This can be done most easily by assigning
numerical weights to each category and calculating the combined score as
S =
N∑
i=1
widi (1)
where S is the combined score, and wi is the weight for the ith category, di is the duration of
the ith category, and N is the total number of categories. Suitable numerical weights must
be chosen for each category; the weights used are shown in table 8. The actual numerical
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
210 Serene Hyun-Jin Choi and Timo A. Nieminen
values are somewhat arbitrary; however, the more desirable categories are (and must be)
assigned higher values.
Table 8. Weights for cognitive and social sub-categories
Cognitive play Social interaction
Sub-categories Weights Sub-categories Weights
Non-play -1 Negative -1
Stereotype 0 No interaction 0
Exploratory 1 Passive-low 1
Functional 2 Passive-high 2
Constructive 2 Unilateral 2
Symbolic 3 Active-low 3
Rule-governed 5 Active-high 5
3. Results
In general, the observed behaviour in children with autism showed much variation over
the sessions, for both cognitive and social categories. This is to be expected—it would be
highly unusual if the children with autism displayed identical behaviour in each session,
given that children with autism have sufficient cognitive abilities to engage in a wide range
of behaviours, despite general intellectual deficits (American Psychiatric Association, 2000;
Sadock & Sadock, 2007). It is possible that children with autism might well show greater
variation in their behaviour than typically developing children due to frequent and unpre-
dictable shifts in attentiveness between on-task and off-task (Hume & Odom, 2007; Pelios,
MacDuff, & Axelrod, 2003). In addition, the play sessions in this study were conducted
over a period of six months, so extraneous confounding factors at home and school beyond
experimental control may well influence the behaviour of not only the children with autism
but also peer players. The play sessions were conducted in naturalistic settings at the par-
ticipating SPSs and SEDUs, without artificial constraints that might restrict the range of
behaviour. All of these factors might have contributed to the observed large variation in
behaviour.
While the observed behaviour was quite variable, this session-to-session variation is ac-
tually of little importance—what is most important are the overall level of social interaction
and the overall level of cognitive play behaviour. In particular, the use of overall scores is
useful for the small sample size in this study to obtain greater statistical power. In this study,
it is the difference in the level of social interaction and cognitive play behaviour between
the pre-training and post-training play sessions that is crucial.
To determine whether or not the difference between the behaviour in the pre-training
and post-training sessions is statistically significant, the mean behaviour and the standard
error in the mean were calculated; this allows, firstly, a simple visual comparison of pre-
training and post-training behaviour on graphs presenting the data, and secondly, testing of
the statistical significance of the change from the pre-training to the post-training behaviour
by determining the 95% confidence intervals for the differences in the means. This is done
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 211
for both the total cognitive play scores and social interaction scores, and for each cognitive
and social category for each dyad.
On figures 5 to 8, the means and the intervals specified by the standard error (that is, the
interval x¯ −∆x¯ to x¯ + ∆x¯) are shown by the grey bars—the bar itself shows the interval
specified by the mean and standard error, and the line through the centre shows the mean. If
the pre-training and post-training gray bars overlap, then there is no statistically significant
difference. if the vertical separation between the bars is large compared to their widths, then
the difference is clearly statistically significant. If the two bars almost touch, then the tables
10 to 24, giving the changes, the standard errors in the changes, the p values, and whether
or not the change is statistically significant at the α = 0.05 level, should be consulted. An
increase that is statistically significant at the α = 0.05 level is indicated by an upwards
arrow ⇑, a statistically significant decrease at α = 0.05 by a downwards arrow ⇓, and a
dash − indicates that the change, if any, was not statistically significant at α = 0.05.
3.1. Indoor play
CA1 CA2 CA3 CA5 All CAs
0
0.5
1
1.5
2
Trained dyads
So
cia
l in
te
ra
ct
io
n
CA1 CA2 CA3 CA5 All CAs
0
0.5
1
1.5
2
Untrained dyads
So
cia
l in
te
ra
ct
io
n
Figure 5. Social interaction behaviour (indoor setting)
Overall, in the indoor settings, the children with autism showed a large increase of
social interaction in trained dyads. The increase was both clearly statistically significant
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
212 Serene Hyun-Jin Choi and Timo A. Nieminen
Table 9. Indoor social interaction
Trained Untrained
CA Pre Post Pre Post
CA1 0.496± 0.208 1.026± 0.099 0.097± 0.043 0.129± 0.034
CA2 0.444± 0.049 0.517± 0.101 0.281± 0.070 0.237± 0.064
CA3 0.211± 0.056 0.448± 0.060 0.303± 0.069 0.150± 0.043
CA5 0.766± 0.252 1.619± 0.289 0.275± 0.081 0.437± 0.070
all 0.479± 0.109 0.902± 0.085 0.239± 0.039 0.238± 0.032
Table 10. Change in indoor social interaction
Trained Untrained
CA Change p Change p
CA1 0.530± 0.200 0.020 ⇑ 0.032± 0.067 0.875 −
CA2 0.073± 0.181 0.695 − −0.044± 0.120 0.813 −
CA3 0.237± 0.111 0.050 − −0.153± 0.077 0.204 −
CA5 0.853± 0.427 0.062 − 0.162± 0.115 0.708 −
all 0.423± 0.129 0.002 ⇑ −0.001± 0.047 0.991 −
Table 11. Indoor social interaction
Trained Untrained
Cat. Pre Post Pre Post
NEG 0.004± 0.001 0.053± 0.011 0.001± 0.001 0.006± 0.004
NOI 0.656± 0.039 0.343± 0.025 0.781± 0.029 0.782± 0.029
P-L 0.278± 0.026 0.380± 0.025 0.201± 0.025 0.186± 0.026
P-H 0.026± 0.009 0.153± 0.016 0.002± 0.003 0.002± 0.002
UNI 0.004± 0.004 0.001± 0.000 0.012± 0.015 0.019± 0.007
A-L 0.009± 0.004 0.041± 0.012 0.004± 0.002 0.006± 0.002
A-H 0.024± 0.017 0.029± 0.015 0.000± 0.000 0.000± 0.000
Table 12. Change in indoor social interaction
Trained Untrained
Cat. Change p Change p
NEG 0.049± 0.002 0.000 ⇑ 0.005± 0.001 0.000 ⇑
NOI −0.313± 0.046 0.000 ⇓ 0.001± 0.035 0.982 −
P-L 0.102± 0.031 0.002 ⇑ −0.015± 0.030 0.615 −
P-H 0.127± 0.010 0.000 ⇑ −0.001± 0.004 0.898 −
UNI −0.003± 0.004 0.478 − 0.007± 0.018 0.683 −
A-L 0.032± 0.005 0.000 ⇑ 0.002± 0.003 0.416 −
A-H 0.005± 0.021 0.818 − 0.000± 0.000 1.000 −
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 213
Table 13. Indoor cognitive play
Trained Untrained
CA Pre Post Pre Post
CA1 0.286± 0.102 0.709± 0.079 0.033± 0.027 0.168± 0.084
CA2 0.591± 0.134 0.775± 0.128 0.490± 0.069 0.470± 0.107
CA3 −0.131± 0.114 0.094± 0.072 −0.001± 0.162 −0.382± 0.056
CA5 1.735± 0.139 1.881± 0.159 1.680± 0.114 1.103± 0.118
all 0.620± 0.068 0.864± 0.059 0.550± 0.059 0.340± 0.056
Table 14. Change in indoor cognitive play
Trained Untrained
CA Change p Change p
CA1 0.423± 0.133 0.007 ⇑ 0.135± 0.155 0.328 −
CA2 0.184± 0.238 0.453 − −0.020± 0.193 0.934 −
CA3 0.225± 0.141 0.133 − −0.382± 0.146 0.027 ⇓
CA5 0.145± 0.235 0.545 − −0.577± 0.186 0.024 ⇓
all 0.244± 0.080 0.004 ⇑ −0.211± 0.071 0.004 ⇓
Table 15. Indoor cognitive play
Trained Untrained
Cat. Pre Post Pre Post
NON 0.492± 0.041 0.481± 0.026 0.553± 0.036 0.583± 0.033
STE 0.081± 0.020 0.049± 0.012 0.079± 0.025 0.134± 0.012
EXP 0.207± 0.038 0.139± 0.018 0.161± 0.021 0.139± 0.024
FUN 0.151± 0.037 0.201± 0.023 0.143± 0.036 0.093± 0.016
CON 0.013± 0.014 0.050± 0.026 0.008± 0.009 0.005± 0.007
SYM 0.055± 0.034 0.066± 0.021 0.056± 0.025 0.046± 0.013
R-G 0.000± 0.000 0.015± 0.008 0.000± 0.000 0.000± 0.000
Table 16. Change in indoor cognitive play
Trained Untrained
Cat. Change p Change p
NON −0.011± 0.048 0.819 − 0.030± 0.043 0.488 −
STE −0.033± 0.023 0.165 − 0.054± 0.031 0.082 −
EXP −0.069± 0.046 0.138 − −0.022± 0.026 0.394 −
FUN 0.050± 0.044 0.258 − −0.050± 0.044 0.261 −
CON 0.037± 0.016 0.028 ⇑ −0.003± 0.011 0.774 −
SYM 0.011± 0.040 0.790 − −0.010± 0.031 0.756 −
R-G 0.015± 0.000 0.000 ⇑ 0.000± 0.000 1.000 −
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
214 Serene Hyun-Jin Choi and Timo A. Nieminen
CA1 CA2 CA3 CA5 All CAs
−1
0
1
2
3
Trained dyads
Co
gn
itiv
e
pl
ay
CA1 CA2 CA3 CA5 All CAs
−1
0
1
2
3
Untrained dyads
Co
gn
itiv
e
pl
ay
Figure 6. Cognitive play behaviour (indoor setting)
(p = 0.002) and large, with the mean social interaction score approximately doubling. In
the untrained dyads, on the other hand, the was no change, with the predominant feature
of the indoor social behaviour shown by children with autism in untrained dyads being
uniformity.
This clearly demonstrates the success of the peer training, and the lack of both measur-
able learning how to socially interact with children with autism by the untrained peers and
any crossover effect in social interaction of children with autism.
Interestingly, negative interaction was significantly increased in both trained and un-
trained dyads (table 12). While the total duration of negative behaviour was still small, its
degree of increase was surprisingly larger in trained dyads than in untrained dyads. In the
trained dyads, this increase was counterbalanced by a large decrease in no interaction, and
increases in appropriate social behaviour. This increase did not appear to be a problem,
and at least partly represents failed attempts by the trained peers to initiate to and interact
with the children with autism. Similar increases in negative behaviour have been noted by
others. Guralnick, Connor, Hammond, Gottman, and Kinnish (1996) reported such an in-
crease, and assumed that it was due to the higher activity levels in inclusive setting. In the
study presented in this paper, it was observed that when trained peers started to interact,
children with autism resisted against the interaction. “Being alone” time (no-interaction be-
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 215
haviour) of the children with autism was constantly interrupted in the trained dyads because
trained peers kept initiating and responding to play together (“desirable” interruption by the
trained peer) despite resistance (negative interaction) from the children with autism.
Occasionally, this negative interaction in children with autism was also caused by “im-
mature” interaction delivered from trained peers. Although trained peers have learnt how
to interact with the children with autism, it did not mean they always know how to handle
uncooperative interaction such as rejecting, ignoring, and avoiding behaviours shown by
children with autism. They did not seem to know when was the best time to interact again
after strong resistance from children with autism. Sometimes, trained peers missed the time
when the children with autism were ready, or could not notice the best time to interact again.
This indicates that one more component—observing and waiting skill—for play interaction
skills may be required in the peer training program.
In spite of the difficulties in interacting with the children with autism, trained peers were
sufficiently enthusiastic so as to keep interacting with them. With this consistent interac-
tion provided by trained peers, the children with autism significantly increased their social
interactions by showing appropriate responses (passive-low and passive-high interactions).
Initiation (active-low interaction) in children with autism was also significantly increased
by positive or non-negative responses from trained peers. These results, increases of social
interaction in children with autism by interacting with trained peers, support other studies
on peer-mediated interventions (e.g., Harper, Symon, & Frea, 2008).
With the untrained peers, there was almost no change in the pattern of social interac-
tions in the children with autism other than the increase in negative interaction noted above.
Essentially, and especially when compared to the intensive interaction trials shown by the
trained peers as the result of peer training, untrained peers did not interact with children
with autism. It does not mean that untrained peers did not try to play with or interact with
the children with autism—untrained peers occasionally showed some general social con-
ventional manners such as greeting, smiling, suggesting sandpit play, providing toys, and
so on—but when untrained peers received no response or unfamiliar interaction behaviour
from children with autism, they showed fewer and fewer further attempts to do so.
Contrasting the result on social interaction in children with autism when paired with
trained peers versus untrained peers, the importance or the necessity of peer training is
clearly evident. Without play training that includes proper information about autism, prac-
tical interaction skills, and so on, these naïve (untrained) peers would not try to interact with
children with autism because they cannot understand the social responses and cues shown
by the children with autism. In addition, uncorrected stereotypical notions or prejudice
about children with disabilities could also contribute to this unwillingness to interact.
If the cognitive level of play were to be used as the primary measure of success of
the training program, the same conclusion would have been reached, as a similar pattern
emerged. However, the increase in cognitive play in trained dyads, while still statistically
significant, was smaller than the increase in social interaction in the same dyads. In the
untrained dyads, there was a statistically significant (but small) decrease in untrained dyads,
as opposed to the absence of change in social interaction in untrained dyads. Some further
discussion is warranted, but this is best deferred until the outdoor play results have been
presented.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
216 Serene Hyun-Jin Choi and Timo A. Nieminen
3.2. Outdoor play
CA1 CA2 CA3 CA5 All CAs
0
0.5
1
1.5
2
Trained dyads
So
cia
l in
te
ra
ct
io
n
CA1 CA2 CA3 CA5 All CAs
0
0.5
1
1.5
2
Untrained dyads
So
cia
l in
te
ra
ct
io
n
Figure 7. Social interaction behaviour (outdoor setting)
The overall pattern of change in social interaction in outdoor settings was very similar
to that seen in indoor settings—a large and statistically significant improvement in trained
dyads, and no change in untrained dyads. For the trained dyads, this increase was even
larger than in the indoor settings, with a post-training social interaction score over three
times larger than the pre-training score.
This even larger increase may have resulted from the playground facilities and apparatus
being more familiar and age-appropriate for the peers than the indoor setting toys (which
were chosen to be suitable for the children with autism). The peers could then concentrate
on social interaction to play with children with autism in outdoor setting, rather than being
distracted by novel toys. The children with autism may have also been interested in learning
how to play with new equipment, and been more willing to listen to and follow the peers.
However, compared to the indoor settings, a smaller proportion of time was spent in
social interaction. This is likely to be due to the wide-open setting in the playground—it
was easier for the two children to play without interaction, and if the children with autism
wished to avoid interaction, they had much more space and opportunity to do so. The
availability of such easy “escape” from unwanted interaction may explain the lower level
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 217
Table 17. Outdoor social interaction
Trained Untrained
CA Pre Post Pre Post
CA1 0.132± 0.060 1.069± 0.181 0.101± 0.089 0.055± 0.016
CA2 0.199± 0.071 0.428± 0.083 0.231± 0.107 0.186± 0.050
CA3 0.022± 0.009 0.507± 0.146 0.021± 0.012 0.022± 0.004
CA5 0.431± 0.126 0.872± 0.145 0.176± 0.032 0.138± 0.033
all 0.196± 0.046 0.719± 0.069 0.132± 0.029 0.100± 0.020
Table 18. Change in outdoor social interaction
Trained Untrained
CA Change p Change p
CA1 0.937± 0.204 0.001 ⇑ −0.046± 0.058 0.825 −
CA2 0.230± 0.152 0.154 − −0.045± 0.103 0.772 −
CA3 0.484± 0.236 0.062 − 0.001± 0.015 0.997 −
CA5 0.441± 0.218 0.062 − −0.038± 0.053 0.865 −
all 0.523± 0.056 0.000 ⇑ −0.032± 0.038 0.398 −
Table 19. Outdoor social interaction
Trained Untrained
Cat. Pre Post Pre Post
NEG 0.004± 0.002 0.007± 0.002 0.008± 0.006 0.000± 0.000
NOI 0.817± 0.037 0.492± 0.040 0.851± 0.029 0.901± 0.020
P-L 0.162± 0.031 0.289± 0.029 0.141± 0.029 0.098± 0.019
P-H 0.016± 0.011 0.204± 0.030 0.000± 0.000 0.000± 0.000
UNI 0.000± 0.000 0.000± 0.000 0.000± 0.000 0.000± 0.000
A-L 0.002± 0.003 0.008± 0.004 0.000± 0.000 0.000± 0.001
A-H 0.000± 0.000 0.001± 0.001 0.000± 0.000 0.000± 0.000
Table 20. Change in outdoor social interaction
Trained Untrained
Cat. Change p Change p
NEG 0.003± 0.003 0.308 − −0.008± 0.008 0.292 −
NOI −0.325± 0.046 0.000 ⇓ 0.049± 0.037 0.189 −
P-L 0.127± 0.038 0.002 ⇑ −0.042± 0.037 0.251 −
P-H 0.189± 0.013 0.000 ⇑ 0.000± 0.000 0.000 ⇑
UNI 0.000± 0.000 1.000 − 0.000± 0.000 0.000 ⇑
A-L 0.005± 0.004 0.134 − 0.000± 0.000 0.000 ⇑
A-H 0.001± 0.000 0.000 ⇑ 0.000± 0.000 1.000 −
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
218 Serene Hyun-Jin Choi and Timo A. Nieminen
Table 21. Outdoor cognitive play
Trained Untrained
CA Pre Post Pre Post
CA1 0.382± 0.200 0.822± 0.099 0.245± 0.110 0.347± 0.101
CA2 0.486± 0.181 0.953± 0.205 0.443± 0.093 1.033± 0.228
CA3 −0.063± 0.035 0.376± 0.154 −0.003± 0.043 0.251± 0.325
CA5 1.275± 0.089 1.563± 0.070 1.178± 0.113 1.439± 0.096
all 0.520± 0.077 0.928± 0.079 0.466± 0.047 0.767± 0.091
Table 22. Change in outdoor cognitive play
Trained Untrained
CA Change p Change p
CA1 0.440± 0.212 0.063 − 0.102± 0.176 0.640 −
CA2 0.466± 0.377 0.236 − 0.590± 0.374 0.143 −
CA3 0.439± 0.250 0.105 − 0.254± 0.262 0.340 −
CA5 0.288± 0.115 0.025 ⇑ 0.260± 0.161 0.037 ⇑
all 0.408± 0.095 0.000 ⇑ 0.302± 0.060 0.000 ⇑
Table 23. Outdoor cognitive play
Trained Untrained
Cat. Pre Post Pre Post
NON 0.657± 0.050 0.454± 0.037 0.703± 0.030 0.473± 0.050
STE 0.034± 0.017 0.029± 0.010 0.019± 0.010 0.066± 0.018
EXP 0.064± 0.015 0.078± 0.019 0.072± 0.020 0.089± 0.023
FUN 0.245± 0.034 0.437± 0.042 0.206± 0.022 0.372± 0.041
CON 0.000± 0.000 0.003± 0.003 0.000± 0.000 0.000± 0.000
SYM 0.000± 0.000 0.000± 0.000 0.000± 0.000 0.000± 0.000
R-G 0.000± 0.000 0.000± 0.000 0.000± 0.000 0.000± 0.000
Table 24. Change in outdoor cognitive play
Trained Untrained
Cat. Change p Change p
NON −0.203± 0.062 0.002 ⇓ −0.230± 0.039 0.000 ⇓
STE −0.005± 0.020 0.798 − 0.047± 0.013 0.001 ⇑
EXP 0.014± 0.019 0.453 − 0.017± 0.025 0.506 −
FUN 0.192± 0.041 0.000 ⇑ 0.166± 0.029 0.000 ⇑
CON 0.003± 0.000 0.000 ⇑ 0.000± 0.000 1.000 −
SYM 0.000± 0.000 1.000 − −0.000± 0.000 0.534 −
R-G 0.000± 0.000 1.000 − 0.000± 0.000 1.000 −
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 219
CA1 CA2 CA3 CA5 All CAs
−1
0
1
2
3
Trained dyads
Co
gn
itiv
e
pl
ay
CA1 CA2 CA3 CA5 All CAs
−1
0
1
2
3
Untrained dyads
Co
gn
itiv
e
pl
ay
Figure 8. Cognitive play behaviour (outdoor setting)
of negative interaction compared to the indoor settings. Generally, it was more difficult
for trained peers to establish and maintain social interaction with children with autism.
However, the post-training social interaction score for the trained dyads was almost as high
as in the indoor settings, despite the pre-training score being much lower, so the trained
peers were clearly able to overcome such difficulties.
This increased area in which to avoid one’s play partner is also likely to be the cause of
the lower level of interaction seen in the untrained dyads—avoidance could appear to be a
“safe” strategy for both the children with autism and the untrained peers.
In the trained dyads, three sub-categories (passive-low, passive-high, and active-high
interaction) of social interaction showed statistically significant increases, accompanied by
a corresponding large and statistically significant decrease of no-interaction (see table 20).
The children with autism and trained peers were able to interact in a simple and clear man-
ner in outdoor setting. For example, despite occasional distractions such as PE sessions held
nearby in the playground, children with autism successfully using playground equipment
and sandpit toys with the trained peers. The trained peers then guided and encouraged the
children with autism to try another equipment (sometimes, as mentioned earier, perhaps to
excess).
While statistically significant changes were seen in some categories of social interaction
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
220 Serene Hyun-Jin Choi and Timo A. Nieminen
in untrained dyads, both the pre-training and post-training durations, and the change in
them were less than 0.0005, and only made a negligible contribution to the overall social
behaviour. These changes cannot be regarded as important.
Measurement of the cognitive level of play, on the other hand, provided a surprising
and remarkable result: all children with autism, in both trained dyads and untrained dyads,
showed a similar and large increase in cognitive play in outdoor setting. While the increase
in trained dyads was about 25% greater, this is within the range of variation indicated by
the standard errors in the changes.
The increase in the cognitive level of play shown in the trained dyads was even larger in
the outdoor settings than in the indoor settings. This larger improvement of cognitive play
might have been due to the initial novelty of outdoor playground facilities and apparatus,
compared to the more familiar indoor toys. Since the outdoor setting was located in the
regular state primary school (SPS-A) for CA1, CA2 and CA3, the change might indicate
learning to use these unfamiliar playground facilities. Overall, children with autism seemed
to be quite interested in the playground facilities.
While trained peers interacted with children with autism using the “play interaction
skills” that they learned during peer training, untrained peers mainly played by themselves.
Occasionally, especially at the beginning of the pre-training play phase (i.e., the earliest
play sessions), untrained peers asked children with autism to use playground equipment
together. However, when they could not get appropriate responses (e.g., no response) from
the children with autism, they simply played by themselves. Sometimes, untrained peers
and children with autism were far apart from each other, with, for example, the untrained
peer in the sandpit area and the child with autism using other playground equipment. Un-
trained peers also seemed to be embarrassed when children with autism suddenly initiated
without verbalization—in these kinds of “awkward” situations, untrained peers were not
likely to know how to interact with children with autism, and chose the easiest path for
them: leaving the area or ignoring the child with autism.
Therefore, the improvement in untrained dyads does not appear to result from any im-
provement in social interaction with the untrained peers. It could possibly be due to the
children with autism having been taught by the trained peers how to use the available play-
ground facilities and equipment; since the improvement occurred after the peer-training
phase, this is certainly a possible explanation, and would demonstrate a generalization of
cognitive play skills. However, since the “higher” categories of cognitive play (i.e., sym-
bolic and rule-governed play) were almost completely absent, the improvement in outdoor
cognitive play does not represent any significant cognitive gains by the children with autism.
Since the improvement could be due to this possible carryover effect, or could be due to
the children with autism learning from experience with the facilities and equipment, without
assistance from the peers, little can be concluded.
As noted above, these results were surprising and unexpected. It appears that there
is no primary cognitive deficit in children with autism preventing appropriate use of out-
door playground facilities or equipment, and such play behaviour can be readily learned by
children with autism.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 221
4. Discussion
It was demonstrated clearly and conclusively that when typically developing peers were
systematically trained (peer training), they were able to more effectively and consistently
able to play and interact with children with autism. Untrained peers, on the other hand,
were not able to interact effectively with the children with autism, and the level of social
interaction did not improve over time, indicating that practice alone is ineffective.
The main results found here in this peer-mediated play intervention suggest that social
skill training for only children with autism may have little or no effect on the facilitation
of positive interaction with peers, in the absence of training of peers. Although children
with autism can improve their social skills when carefully trained by educators, this does
not mean that their expression of these social skills is such that typically developing peers
generally understand or accept—their social behaviour is still often regarded as bizarre and
incomprehensible by untrained peers.
Through systematic peer training, typically developing peers are able to realize diversity
in people, and recognize abilities, rather than only disabilities, in children with disabilities
including autism. Trained peers can appreciate the similarities between themselves and
children with autism as well as their differences. They can learn how to play and interact
with children with autism.
Without systematic peer training, typically developing peers may not know how to play
or interact with children with autism, and cannot easily learn to do so through practice
alone. With untrained peers, social development as a result of social interaction in children
with autism cannot be expected.
Therefore, systematic interaction training should be provided not only to children with
autism but also to typically developing peers. When both groups of children are trained,
the optimal outcome for positive social interaction between them results because social
interaction can be maintained in a reciprocal way with initiation followed by appropriate
response in human relationships. With no, or insufficient, initiation, then response will not
occur or will be infrequenct. If interaction is initiated, but the initiation is not recognized
by the recipient, no response occurs. If the initiation is perceived negatively, the response is
likely to be negative. Such failures of interaction were observed in this study. For example,
when CA1 did not get any response from UP12, despite several attempts, CA1 stopped
trying to interact with UP12. Untrained peers also showed a similar pattern of behaviour,
with unsuccessful attempts at initiation followed by a cessation of such attempts; P7 stated
that when he could not any response from CA3, he stopped greeting CA3 because P7 had
no reason to continue initiating to CA3.
Furthermore, given that social interaction is reciprocal, individualistic, and relative
(Rubin, Bukowski, & Parker, 1998), aggressive behaviour in children with autism should be
examined in the context of reciprocal interaction, investigating environmental factors and
the circumstances leading up to the violent display. For example, in this study, from time
to time CA5 invited UP14 to join in his play activities, but UP14 kept ignoring or rejecting
him without any explanation. Finally, CA5 used aggressive behaviour to express his anger,
throwing toys towards UP14 in the last play session. The reasons why P14 kept ignoring or
rejecting CA5’s invitations are not known. According to her class teacher, she was some-
what shy but compliant, getting along with her classmates. Does this mean that P14 should
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
222 Serene Hyun-Jin Choi and Timo A. Nieminen
have had no difficulty to play with other children, and that the lack of successful interaction
is solely due to CA5 having no appropriate social skills to interact with typically develop-
ing peers? Would it be reasonable to describe CA5 as an aggressive boy with autism? The
answers to both questions are clearly no, as evidenced by CA5 showing positive interaction
and successfully playing with the trained peer, TP9. Therefore, social behaviour in children
with autism should be interpreted from the viewpoint of reciprocal human interaction.
Finally, the surprising and contradictory results obtained from measurement of the cog-
nitive level of play merit some further discussion. While the results from the indoor settings
largely agreed with those deduced from the measurement of the level of social interaction,
the outdoor results were remarkably different, with similar large and statistically signifi-
cant increases in the cognitive level of play being seen in both trained and untrained dyads.
Coincidence of the increase in both groups with the training of the trained peers makes a
carryover effect from the trained dyads to the untrained dyads appear to be the most likely
cause, especially in the presence of an accompanying decrease in social interaction in the
untrained dyads. Since no large cognitive improvement in the children with autism would
be expected in a short-term study (Luckett, Bundy, & Roberts, 2007) such as is reported
here, it is most unlikely that this increase represents any real improvement in cognitive
ability, any more than the small but statistically significant decrease in the cognitive level
of play in untrained dyads in indoor settings indicated a decrease in cognitive ability (with
a simultaneous increase with trained peers!).
Therefore, we strongly urge a degree of caution in the interpretation of results based on
the measurement of the cognitive level of play based on the Piaget–Smilansky categories, or
similar categories, especially for children with autism where the primary deficits are social
rather than cognitive. Possible discrepancies between the cognitive level displayed during
play and the actual level of cognitive ability should be kept in mind.
References
Allodi, M. W. (2000). Self-concept in children receiving special support at school. Euro-
pean Journal of Special Needs Education , 15(1), 69-78.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders: DSM-IV-TR (4th, with text revision, ed.). Washington, DC: American
Psychiatric Association.
Anderson, A., Moore, D. W., Godfrey, R., & Fletcher-Flinn, C. M. (2004). Social skills
assessment of children with autism in free-play situations. Autism, 8(4), 369-385.
Brophy, K., & Zukowski, D. S. (1984). Social and play behaviour of special needs and
non-special needs toddlers. Early Child Development and Care , 13(2), 137-154.
Choi, S. H.-J. (2005). Peer training for improved social interaction in children with autism .
Unpublished doctoral dissertation, The University of Queensland, Brisbane, QLD,
Australia.
Choi, S. H.-J. (2007, December). Peer training methods for children and adolescents with
autism: A review. International Journal of Pedagogies and Learning , 3(3), 92-100.
Farran, D. C., & Son-Yarbrough, W. (2001). Title I funded preschools as a developmental
context for children’s play and verbal behaviors. Early Childhood Research Quar-
terly, 16(2), 245-262.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
Trained Peers’ Interactions With Children With Autism During Play 223
Fox, J. E. (1996). Back-to-basics: Play in early childhood. Early Childhood News, 8(5),
19-24.
Guralnick, M. J., Connor, R. T., Hammond, M., Gottman, J. M., & Kinnish, K. (1996).
Immediate effects of mainstreamed settings on the social interactions and social in-
tegration of preschool children. American Journal on Mental Retardation , 100(4),
359-377.
Harper, C. B., Symon, J. B. G., & Frea, W. D. (2008). Recess is time-in: Using peers to
improve social skills of children with autism. Journal of Autism and Developmental
Disorders, 38(5), 815-826.
Hartup, W. W. (1999). Peer experience and its developmental significance. In M. Ben-
nett (Ed.), Developmental psychology: Achievements and prospects (p. 106-125).
Philadelphia, PA: Psychology Press.
Hetherington, E. M., Parke, R. D., & Locke, V. O. (1999). Child psychology: A contempo-
rary viewpoint (5th ed.). New York, NY: McGraw-Hill.
Hume, K., & Odom, S. (2007). Effects of an individual work system on the independent
functioning of students with autism. Journal of Autism and Developmental Disorders ,
37(6), 1166-1180.
Jackson, J. N., & Campbell, J. M. (2008). Teachers’ peer buddy selections for children
with autism: Social characteristics and relationship with peer nominations. Journal
of Autism and Developmental Disorders , 38, Online first, DOI 10.1007/s10803-008-
0623-1.
Jahr, E., Eikeseth, S., Eldevik, S., & Aase, H. (2007). Frequency and latency of social in-
teraction in an inclusive kindergarten setting: A comparison between typical children
and children with autism. Autism, 11(4), 349-363.
Johnson, J. E., Christie, J. F., & Yawkey, T. D. (1999). Play and early childhood develop-
ment (2nd ed.). New York, NY: Addison Wesley Longman.
Liebal, K., Colombi, C., Rogers, S. J., Warneken, F., & Tomasello, M. (2008). Helping and
cooperation in children with autism. Journal of Autism and Developmental Disorders ,
38(2), 224-238.
Luckett, T., Bundy, A., & Roberts, J. (2007). Do behavioural approaches teach children
with autism to play or are they pretending? Autism, 11(4), 365-388.
Marcus, B. A., Vollmer, T. R., Swanson, V., Roane, H. R., & Ringdahl, J. E. (2001). An
experimental analysis of aggression. Behavior Modification , 25(2), 189-213.
McConnell, S. R. (2002). Interventions to facilitate social interaction for young children
with autism: Review of available research and recommendations for educational in-
tervention and future research. Journal of Autism and Developmental Disorders ,
32(5), 351-372.
Nieminen, T. A., & Choi, S. H.-J. (2008, April). Binary keys for classification and tax-
onomy of behaviour. International Journal of Research and Method in Education ,
31(1), 31-44.
Parten, M. B. (1932). Social participation among pre-school children. Journal of Abnormal
and Social Psychology , 27, 243-269.
Pelios, L. V., MacDuff, G. S., & Axelrod, S. (2003). The effects of a treatment package
in establishing independent academic work skills in children with autism. Education
and Treatment of Children , 26(1), 1-21.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.
224 Serene Hyun-Jin Choi and Timo A. Nieminen
Piaget, J. (1952). The origins of intelligence in children . Oxford, UK: International Uni-
versities Press.
Piaget, J. (1959). The language and thought of the child (3rd ed.). London: Routledge &
Kegan Paul.
Piaget, J. (1962). Play, dreams and imitation in childhood . London: Routledge & Kegan
Paul.
Piaget, J. (1977). The development of thought: Equilibration of cognitive structures . Ox-
ford, UK: Viking.
Pierce, K., & Schreibman, L. (1995). Increasing complex social behaviors in children with
autism: Effects of peer-implemented pivotal response training. Journal of Applied
Behavior Analysis, 28(3), 285-295.
Pink, B. (2008). An introduction to socio-economic indexes for areas (SEIFA): 2006 .
Canberra, ACT, Australia: Australian Bureau of Statistics.
Pugmire-Stoy, M. C. (1992). Spontaneous play in early childhood . Albany, NY: Delmar
Publishers.
Rubin, K. H., Bukowski, W., & Parker, J. G. (1998). Peer interactions, relationship, and
groups. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology (5th ed.,
Vol. 3, p. 619-700). New York: John Wiley.
Sadock, B. J., & Sadock, V. A. (2007). Kaplan and Sadock’s synopsis of psychiatry: Behav-
ioral sciences/clinical psychiatry (10th ed.). Philadelphia, PA: Lippincott Williams
& Wilkins.
Saracho, O. N. (1993). A factor analysis of young children’s play. Early Child Development
and Care, 84, 91-102.
Smilansky, S. (1968). The effects of sociodramatic play on disadvantaged preschool chil-
dren. Oxford, UK: Wiley.
Vygotsky, L. S. (1962). Thought and language . Oxford, UK: Wiley.
Yang, T.-r., Wolfberg, P. J., Wu, S.-c., & Hwu, P.-y. (2003). Supporting children on the
autism spectrum in peer play at home and school: Piloting the integrated play groups
model in Taiwan. Autism, 7(4), 437-453.
Special Education in the 21st Century, Nova Science Publishers, Incorporated, 2009. ProQuest Ebook Central,
Co
py
rig
ht
©
2
00
9.
N
ov
a
Sc
ie
nc
e
Pu
bl
ish
er
s,
In
co
rp
or
at
ed
. A
ll r
ig
ht
s
re
se
rv
ed
.

学霸联盟