Professional Documents
Culture Documents
397.full Articulo Humprey Integracion
397.full Articulo Humprey Integracion
http://aut.sagepub.com/
Published by:
http://www.sagepublications.com
On behalf of:
Subscriptions: http://aut.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://aut.sagepub.com/content/15/4/397.refs.html
What is This?
Downloaded from aut.sagepub.com by guest on March 14, 2012
autism © 2011
Peer interaction patterns SAGE Publications
and The National
Autistic Society
among adolescents with Vol 15(4) 397–419; 387804
1362-3613(2011)
398
399
Method3
Design
A quasi-experimental design was adopted. The independent variable was
SEN group status (ASD, DYS (dyslexia), or CON (no identified special edu-
cational needs)). The dependent variables were type, frequency, and dura-
tion of peer interaction.
Participants
Our sample comprised 111 students4 (38 in the ASD group, 35 in the DYS
group, and 38 in the CON group) drawn from 12 secondary mainstream
schools in the north-west of England (102 males, 9 females; mean age
13y9m)5. This exceeded a target sample size of n = 30 in each group that
would be needed to detect a large effect size at power = 0.80 and alpha =
0.05 (Cohen, 1992). A matched-triad process was undertaken to ensure
that the groups were matched as closely as possible by age and gender. A
one-way analysis of variance (ANOVA) confirmed no statistically signifi-
cant differences between the groups in age (p > .05). Furthermore, a chi-
square test confirmed no statistically significant association between group
and gender (p > .05). Students in the ASD and DYS groups all had relevant
confirmed diagnoses and were on their school’s SEN register. In all cases
diagnoses had been made by an appropriately qualified professional (e.g.
psychiatrist, psychologist) using established assessment techniques (e.g.
Autism Diagnostic Observation Schedule). A matched-pairs process was
undertaken in an attempt to ensure that these two groups were matched
by stage of the SEN code of practice (DfES, 2001 – see footnote 2 above
for explanation of these stages). A chi-square test confirmed no statistically
significant association between group and SEN stage (p > .05). Table 1
provides a brief overview of the demographic information for each group.
Consent was sought and received from the head-teacher at each school,
from each participating student’s parents, and from the students themselves
(see ‘Ethical considerations’ below).
403
Mean age 13y8m (SD = 1.54) 13y9m (SD = 1.57) 13y8m (SD = 1.53)
Gender
Male, % 90 92 90
Female, % 10 8 10
Level of SEN provision
Statement of SEN, % 60 43 N/A
School action plus, % 32 43
School action, % 8 14
Materials
Peer Interaction Observation Schedule (PIOS) The PIOS is a measure
of the type, duration (how long each behaviour lasts) and frequency (how
often a behaviour is observed) of the interaction between a focal student
and their peers. Adapted from an observation schedule used by Pellegrini
and Bartini (2000), the instrument consists of 22 observable behaviour
codes. 15 of these refer to behaviours exhibited by the focal student (e.g.
‘social initiation by focal student’), while the remaining seven pertain to
the behaviours of their peers (e.g. ‘acceptance of social initiation by focal
student’). A full list of the behaviour codes and their definitions can be
found in Appendix 1.
To use the measure, researchers observe the focal student continuously
for the full length of the observation period. The recording schedule is split
into one-minute intervals to make recording easier, and allow tracking of
how patterns of interactions evolve chronologically (that is, behaviours can
be recorded in the sequence in which they occur). If a behaviour occurs
for a minute or more, the appropriate boxes are marked with an X. If a
behaviour changes during a one-minute interval, the approximate pro-
portions of the old and new behaviour are recorded in seconds (e.g. 40
seconds of co-operative interaction, followed by 20 seconds of solitary
engaged behaviour).
A frequency score for each behaviour is calculated by summing the
distinct episodes that occurred in that code during the observation period,
and dividing that number by the total number of behaviours observed
overall. This gives the frequency score as a proportion of all the distinct
behaviours that were recorded. A duration score for each behaviour is calcu-
lated by summing the time recorded in that code during the observation
period, and dividing this by the overall observation time. This gives the
duration score as a proportion of the total observation duration.
404
Procedure
Students were observed during two to four break and lunchtime sessions,
ranging from approximately 15 to 45 minutes in length, over a two-day
period. A team of eight researchers – including the two authors of this
paper – conducted the observations (although each student was only
observed by one researcher). Students were followed as closely as possible
during the observation periods, but the researchers tried not to directly
interact with them. The mean total observation time was 61 minutes.
Ethical considerations
Standard ethical procedures for educational (British Educational Research
Association, 2004) and psychological research (British Psychological Society,
2004) were followed throughout the study. As mentioned above, consent
was sought and received from the head-teacher at each school, from each
participating student’s parents and from the students themselves. Partici-
pating students were given a clear explanation of the purpose of the study,
and informed that their data would be treated in confidence and that they
would remain anonymous. Their right to withdraw at any point was also
made clear. The school special educational needs co-ordinator (SENCO)
405
Results
Descriptive statistics pertaining to the average total observation duration and
average total number of discrete behaviours observed for each group are
presented in Table 2.
Duration
To explore group differences in duration of behaviours observed, a multi-
variate analysis of variance (MANOVA) was performed. As predicted a main
effect of group (F(1.697) = .572, p < .01, η2 = .286) was revealed. Results
are displayed in Table 3.
Analysis by type revealed that students with ASD spent significantly
more time in solitary activities (unoccupied and engaged) and less time in
co-operative interaction than students in the DSY or CON groups. They
also spent more time engaging in reactive aggression (verbal) than either
control group.
406
FOCAL STUDENT
Solitary (unoccupied) .0296 .0119 .0049 4.307* .074
Solitary (engaged) .2437 .0854 .0697 14.969*** .217
Solitary (onlooker) .0378 .0114 .0187 2.793 .049
Parallel .1654 .1472 .1402 .343 .006
Co-operative (interaction) .4087 .6405 .6463 17.978*** .250
Co-operative (game) .0594 .0290 .0441 .581 .011
Rough/vigorous play .0100 .0232 .0133 2.756 .049
Locomotor .0056 .0272 .0307 1.507 .027
Instrumental aggression (verbal) .0013 .0005 .0002 1.938 .035
Instrumental aggression (physical) .0029 .0012 .0008 .689 .013
Social initiation by focal student .0064 .0043 .0101 1.326 .024
Reactive aggression (verbal) .0011 .0002 .0000 3.344* .058
Reactive aggression (physical) .0022 .0001 .0006 2.280 .041
Submissive/passive .0011 .0010 .0003 .658 .012
Acceptance of social initiation by peer .0050 .0034 .0026 .964 .018
PEER(S)
Social initiation by peer .0072 .0043 .0033 1.699 .031
Instrumental aggression (verbal) .0014 .0005 .0003 1.964 .035
407
*p = <.05, ** p = <.01, *** p = <.001.
AU T I S M15(4)
No significant group differences were found in the duration focal
students in the three groups spent in the following behaviours: solitary
(onlooker), parallel behaviour, co-operative (game), rough/vigorous play,
locomotor, instrumental aggression (both verbal and physical), social initi-
ation by focal student, reactive aggression (physical), submissive/passive
behaviour, and acceptance of social initiation by peers. In terms of peer
behaviour, no significant differences were found in the duration of time
peers of students in the three groups spent in the following behaviours:
social initiation by peer, instrumental aggression (verbal and physical),
reactive aggression (verbal and physical), submissive/passive, and accep-
tance of social initiation by focal student.
Discussion
The main findings of the current study were that included students with
ASD spent more time engaged in solitary behaviours, less time engaged in
co-operative interaction with peers, and more time engaging in reactive
aggression towards peers than either comparison group. In terms of fre-
quency of observed behaviours, similar patterns emerged, but additionally
408
FOCAL STUDENT
Solitary (unoccupied) .0324 0.104 .0054 7.032*** .115
Solitary (engaged) .1586 .0840 .0723 7.335*** .120
Solitary (onlooker) .0440 .0147 .0262 3.085* .054
Parallel .1614 .1743 .1800 .171 .003
Co-operative (interaction) .3409 .4416 .4343 7.438*** .121
Co-operative (game) .0226 .0164 .0263 .486 .009
Rough/vigorous play .0198 .0564 .0378 3.574* .062
Locomotor .0083 .0313 .0292 1.142 .021
Instrumental aggression (verbal) .0082 .0043 .0014 1.878 .034
Instrumental aggression (physical) .0032 .0095 .0041 .623 .011
Social initiation by focal student .0381 .0326 .0463 .695 .013
Reactive aggression (verbal) .0082 .0022 .0004 3.811* .066
Reactive aggression (physical) .0040 .0025 .0101 .710 .013
Submissive/passive .0054 .0045 .0027 .414 .008
Acceptance of social initiation by peer .0316 .0245 .0164 1.795 .032
PEER(S)
Social initiation by peer .0538 .0322 .0212 3.327* .058
Instrumental aggression (verbal) .0086 .0016 .0026 3.139* .055
409
*p = <.05, ***p = <.001
AU T I S M 15(4)
participants with ASD engaged in less instances of rough/vigorous play, and
were subject to more instances of social initiation and instrumental verbal
aggression by peers than either comparison group. These findings broadly
support the REPIM model introduced earlier in this article (although there
is one exception, which is discussed below) and that of other authors, who
have consistently found that such students find interacting with their peers
difficult, and have fewer friends, more limited social networks, and less
peer social support as a result (e.g. Cairns and Cairns, 1994; Humphrey and
Symes, 2010a; Kasari and Rotherham-Fuller, in press). In this section we
address each of our key findings and consider the implications for our
theoretical model and practice in this area.
The finding that included students with ASD spent more time engaged
in solitary behaviours and less time engaged in co-operative interaction with
peers than the two control groups resonates with Bauminger and colleagues’
(2003) finding that such students spent less time engaged in social inter-
action than typically their developing peers. This is also broadly in line with
the conclusions of Lord and Magill-Evans (1995), Hauck et al. (1995), and
Stone and Caro-Martinez (1990) (although as noted before these studies
reported on students in special school settings). The predictions of the
REPIM model are also borne out here. Importantly, students with ASD spent
around 25% less of their time engaged in co-operative interaction and
around 17–20% more of their time engaged in solitary behaviour than the
participants in the two control groups (see Table 3). This obviously places
limits on the opportunities to practise and develop social and communica-
tive skills that are so vital in the development of positive peer relationships.
Another key finding was that peers of students with ASD engaged in
verbal instrumental aggression more frequently than the peers of students
in the other two groups. This is also in line with the REPIM model and
previous research in this area, which demonstrated that these students are
at a higher risk of bullying than other groups of learners (e.g. Humphrey
and Symes, 2010a; National Autistic Society, 2006). The finding could also
help to explain why they are less likely to report that they are being bullied
(National Autistic Society, 2006). If bullying is verbal rather than physical
in nature, the social impairments that characterize ASD may prevent them
from interpreting what is happening to them as bullying (Moore, 2007).
Such a finding has important implications for intervention. Included
students with ASD could be taught, for example, how to recognize when
they are being bullied verbally through the use of social stories (e.g.
Reynhout and Carter, 2006; Rowe, 1999).
The fact that students with ASD were significantly more likely to engage
in reactive verbal aggression than either control group can perhaps be
explained as a direct response to the greater proportion of instrumental
410
Limitations
This study strove to build upon existing knowledge about the social worlds
of included students with ASD and, in particular, to develop understand-
ing of their peer interaction patterns. Observing students matched on a
number of key variables in natural social settings using trained researchers
and a structured observation schedule strengthened the validity of our
findings. However, like all studies, there are some limitations that need to
be considered.
Firstly, although a strength of this study was that it involved the obser-
vation of behaviour in natural settings, the importance of gaining informed
consent from the participants meant that there were always aware they were
being watched – covert observation was not an option here. This will un-
doubtedly have influenced the behaviour of participants. For example, a
minority of students attempted to run away and hide from the researchers,
while others made social approaches. Alongside this, peers of the focal
students may also have been aware that they were being watched and
changed their behaviour accordingly, such as being less likely to engage in
acts of physical aggression towards the focal students (given this, the
increased peer verbal aggression outlined above is all the more striking).
Efforts were made to minimize the influence of the researchers’ presence,
such as observing students from a reasonable distance, but the influence of
this artefact of the study methodology cannot be discounted.
While the PIOS observation schedule used is well established and has
been used in other published studies (e.g. Pellegrini and Bartini, 2000), it
is not completely exhaustive and therefore we do not claim that it is able to
capture the full variety of student peer interaction behaviours. For instance,
the codes used (see Appendix 1) do not capture the ‘micro’ behaviours
(such as eye contact) that may be important in social interaction, particu-
larly in relation to the behaviour of students with ASD (for instance,
Bauminger et al. (2003) found that such students were less likely to make
eye contact during social interaction than their typically developing peers).
That said, the PIOS’s strength is that it captures a broader range of ‘macro’
behaviours than other schedules that have been used in this area, and as
such was considered an ideal instrument for testing the predictions of the
REPIM model.
A further limitation of the study reported in this article was the lack of
independent diagnostic confirmation using a single instrument for parti-
412
Conclusion
The current study is the first comparative study to report on the peer inter-
action patterns of students with ASD in secondary mainstream settings using
observational techniques. Our main findings were that, in terms of dura-
tion, participants with ASD spent more time engaged in solitary behaviours,
less time engaged in co-operative interaction with peers, and more time
engaging in reactive aggression towards peers than either comparison group.
In terms of frequency, similar patterns emerged, but additionally partici-
pants with ASD engaged in less instances of rough/vigorous play, and
were subject to more instances of social initiation and instrumental verbal
aggression by peers than either comparison group. These findings align
well with previous research in this area and provide support for our theo-
retical model of the relationship between peer interaction patterns and
social outcomes for students with ASD (see Figure 1) – although there is
a clear need for this inductive model to be more rigorously tested in a
single, large scale study where the various components can be tested simul-
taneously and subjected to appropriate analysis (such as structural equation
413
Notes
1 Dyslexia is defined as being ‘evident when fluent and accurate word identification
(reading) does not develop, or does so very incompletely’ (British Psychological
Society, 1999).
2 Students identified as having SEN in schools in England are classified according to
the nature and level of additional provision they receive as a result of their
difficulties. Thus, they can be at School Action (SA), School Action Plus (SAP) or in
receipt of a Statement of SEN (SSEN). Students at SA have their special needs met
within their school’s normal resources. Those at SAP are likely to have additional
support from an external agency (e.g. educational psychologist). Finally, students
whose needs have not been met at either SA or SAP will typically undergo a full
statutory assessment of their needs, resulting in the production of an SSEN, which
legally secures a particular level of resources that can be used to support the
student.
3 All of the data reported in this article were collected during the execution of a
larger project on inclusive education for students with ASD funded by the
Economic and Social Research Council (grant reference RES-061-25-0054).
4 Our original sample comprised 40 students in each group (total N = 120), but
lack of school and/or student consent meant that 9 students (2 in the ASD group,
5 in the DYS group and 2 in the CON group) did not participate in the
observational strand of the larger study (see ‘Ethical considerations’).
5 Note: This male:female ratio does not match the aforementioned population
estimates of 3:1/5:1. Aside from being a peculiarity of our sample, this may
indicate that the male:female ratios for ASD incidence in mainstream schools differ
from the population of individuals with ASD.
References
Adler, P.A., Kless, S.J. & Adler, P. (1992) ‘Socialization to Gender Roles: Popularity
among Elementary School Boys and Girls’, Sociology and Education 65: 169–87.
Anderson, A., Moore, D., Godfrey, R. & Fletcher-Flinn, C. (2004) ‘Social Skills
Assessment of Children with Autism in Free-Play Situations’, Autism 8: 1362–3613.
414
415
416
417
418
419