British Journal of Educational Psychology (2000), 70, 337 351 Printed in Great Britain

# 2000 The British Psychological Society

Effects of different forms of school contact on

childrens attitudes toward disabled and nondisabled peers
Pam Maras*
University of Greenwich, London

Rupert Brown
University of Kent at Canterbury

Background. There have been fluctuations in research interest into the

inclusion of children with disabilities in mainstream schools over the last
twenty years. It is still not clear what methods, practices and types of contact
are most likely to promote positive attitudes in children toward disabled peers
and disability generally.
Aims. To consider two theoretical models of inter-group contact, both
claiming to identify precursors for generalised attitude change, in relation to
the attitudes of non-disabled children toward disabled peers as a function of
different classroom contact.
Sample. Participants were 256 non-disabled school children aged 5 11 years
(128 girls and 128 boys).
Methods. Measures of sociometric preference and the evaluation of
psychological and physical attributes were used to ascertain childrens
perceptions of known and unknown peers with disabilities.
Results. A relationship was found between the type of contact the children had
with disabled peers, and their perceptions of psychological and physical
attributes (stereotypes) of groups of unknown disabled and non-disabled
peers.
Conclusions. Results show generalisation of stereotypic attitude/judgments
from one type of disability to another as a consequence of the two types of
contact situation. Findings have important implications for integrating
disabled children into mainstream.

*Requests for reprints should be addressed to Dr Pam Maras, School of Social Science, University of
Greenwich, Avery Hill Campus, Avery Hill Road, London SE9, UK (e-mail: P.F.Maras@gre.ac.uk).

338

Pam Maras and Rupert Brown

The twenty years following the Warnock Report (DES, 1978), subsequent legislation
(e.g., DES, 1981; DfE, 1998, 1993) and codes of practice (e.g., DfE, 1994) saw a
plethora of debate and research on the integration of children with special educational
needs (SENs) into mainstream schools. Controversy encompassed a range of disciplines
and focused on political, social, pedagogical and theoretical issues. Debate
encompassed a range of often competing positions, from the view that prejudice is
socially constructed (e.g., see Barnes, 1996; Oliver, 1990; Shakespeare & Watson, 1997)
to the perspective that disabled people s cognitions about their situations are at the root
of others attitudes toward them (e.g., Johnston, 1997). Since the mid 1990s research
and interest in integration has seemingly waned, the focus having turned more toward
the behaviour of children and young people in and out of school (Maras, 1996). The
descriptor `integration has been superseded in common parlance by `inclusion . These
changes in discourse are reflected in and reflect government papers (e.g. DfEE, 1997,
1998) which, though adopting the term inclusion, link it directly to schools and school
effectiveness (see White & Barber, 1997): a marriage which, it has been suggested, is
fraught with tensions (Lunt & Norwich, 2000).
Notwithstanding these changes in focus, issues in the area of inclusion and attitudes
to disability remain the same and many research questions remain unresolved; it is still
not clear whether, and what form of, contact arising out of inclusive practice produces
positive outcomes in terms of non-disabled children s attitudes. Despite the shift away
from research and moves toward ideological debate in the area, the number of children
and young people with SENs being included in mainstream schools, with the exception
of children with behaviour problems, is seemingly increasing (Norwich, 1990, 1994,
1997). However, theory about the most effective methods of achieving this inclusion is
still relatively undeveloped in relation to the impact of contact on non-disabled (ND)
children s attitudes toward disabled peers with whom they have contact and, more
importantly, toward disabled people generally. From a social psychological viewpoint
increased moves toward inclusion have a direct effect on the amount of social
interaction occurring between mainstream children and children with disabilities and on
their subsequent social perceptions.
There is some evidence on the effects of educational integration on ordinary
children s attitudes (Brinker 1985). In the literature that addresses this issue and that
has looked at children s attitudes towards peers with disabilities in other settings,
several broad underlying themes can be identified. Some research has focused on
contact per se (McConkey, McCormick, & Naughton, 1983), exposure and visibility
(Furnham & Pendred, 1984; Strohmer, Grand, & Purcell, 1984), and perceived
similarities (Siperstein & Chatillon, 1982). The broad conclusion from this work is that
contact per se will have positive outcomes and will be even more effective where
similarities are perceived. Other research has added to the picture by looking at
cognitive mechanisms which might identify why exposure could be an important feature
in attitude formation (Lewis & Lewis 1987). A third strand of research has been
concerned with the effects of contact improving attitudes toward disability (e.g.,
Chesler, 1965), the structure of contact (e.g., Acton & Zarbatany, 1988) and on the
impact of co-operation in improving attitudes (e.g., Armstrong, Johnson, & Balow,
1981; Johnson & Johnson, 1981). However, little is still known about the precise
methods of inclusion which produce the most positive effects and, in particular, how

Childrens attitudes toward disabled peers

339

improved attitudes can be encouraged, not just towards a few individuals, but towards
disabled children in general. In some earlier research we showed how a planned
intervention involving co-operative activities between children with and without
disabilities can generate more favourable intergroup attitudes (Maras & Brown, 1996).
In this paper we further examine this issue by studying children s attitudes in a range of
schools where different inclusive practice is in place. We should note that in this paper
we are not assuming links between children s attitudes and their understanding of
SENs. See Dockrell, Logotheti, and Magiati (1998) for work in this area.
The research reported in this paper utilised two contrasting theoretical perspectives
on contact. For both, the generalisation of attitudes arising out of a contact situation is
a desired outcome. One perspective maintains that for attitudes arising out of contact to
be generalised, references to relevant categories should be de-emphasised we call this
interpersonal, `decategorised contact (Brewer & Miller, 1984). The second proposes
that for positive attitude change arising out of contact to be generalised, some category
salience should be retained and additional features, such as the acknowledgement of
valued differences between groups should be highlighted this we refer to as intergroup, `decategorised contact (Hewstone & Brown, 1986). Brewer and Miller see the
blurring or breaking down of group or category boundaries as essential to
personalisation (and the reduction of categorical biases). Hewstone and Brown
propose that maintaining group boundaries can be beneficial for aiding positive
generalisation since the transfer of positive attitudes from the individuals one has met in
the contact situation to other category members not encountered is thereby facilitated.
In both models, the importance of Allport s (1954) conditions for successful contact
(i.e., co-operative activity between equal status participants with strong institutional
support for the goal of integration) is recognised.
There is some research evidence which supports both perspectives (see Bettencourt,
Charlton, & Kernahan, 1997; Brown, Vivian, & Hewstone, 1999; Maras & Brown, 1996;
Miller, Brewer, & Edwards, 1985; Vivian, Hewstone, & Brown, 1997). However, with the
exception of Maras and Brown (1996), there is little evidence from naturalistic contexts
in support of these models and, again with the exception of Maras and Brown (1996),
both models and much of the research they have instigated have been mainly concerned
with contact between ethnic, national or ad hoc groups. It remains to be seen how
applicable they are to social situations involving children with and without disabilities.
The research reported in this paper tested the appropriateness of Hewstone and
Brown s (categorised) and Brewer and Millers (decategorised) theories of contact as
models for improving attitudes within the social context of inclusive education for
disabled children. It was hypothesised that the former theoretical stance would be the
most conducive to generalised attitude change. It should be noted, however, that we did
not preclude the notion that in `categorised contact situations where the additional
features of institutional support and valued differences are not present, resulting
attitudes may well still be generalised but may not be positive.
Context
The research was conducted in a large local education authority (LEA) in Southeast
England. An extensive preliminary survey identified models of integration effective
within the LEA that related directly to the two models of contact (Maras, 1993):

340

Pam Maras and Rupert Brown

1. Where integration was taking place but where children with disabilities were not
clearly identified by the schools to their mainstream peers as being members of a wider
group (interpersonal `decategorised contact).
2. Where integration was occurring but where children with disabilities were clearly
identified by the schools as members of a group of similar others (inter-group
`categorised contact). This identification took several forms. For example, teachers
talked about disability to the children and the disabled children were taught separately
for all or part of the time. Further examples of this can be seen below in Table 1.
Within these models children with three types of disability were included: hearing
impairment (HI), learning disabilities (LD) and physical disabilities (PD). Schools were
also identified where no obvious inclusion of disabled children (ND) was taking place;
two of these served as control schools. The identification of categorisation, or not (type
of contact) and the matching of schools was crucial to the design of the reported study.
Initially type of contact was assessed by looking at the visibility of children with
disabilities, along with criteria such as school size, significant numbers and spread
across classes of children with significant SENs. Categorisation in terms of withdrawal,
adult helpers, physical location of children with SENs (i.e., in mainstream or separate
classrooms) and the use of aids, such as body-worn hearing aids in the case of children
with HI, were considered. This involved both the survey to class and head teachers
described above and follow-up visits to specific schools. As part of this process both
teachers and headteachers along with other appropriate adults such as SENs teachers
were interviewed, observations were carried out and at least two further follow-up visits
were made to schools thought to fit the models. Eight schools were finally identified
and took part in the study. Only schools where one type of disability (LD, HI, PD or
none) predominated participated in the study. Thus there were two schools representing
each type of disability inclusion. Table 1 provides examples of how the two types of
contact (`categorised and `decategorised ) were operationalised in the schools where HI
children were integrated.

Table 1. Features of categorised and decategorised schools where children who are
hearing impaired are included
CATEGORISED CONTACT

DECATEGORISED CONTACT

Group and individual differences and similarities are acknowledged.

HI unit is in a separate building from the
mainstream school. Reverse integration takes
place in a structured way.
HI children are encouraged to wear `bodyworn hearing aids.
HI is discussed formally with mainstream
children & they are given information about
how to communicate with HI children.

Only individual differences and similarities are

acknowledged.
HI unit is in mainstream school. Reverse
integration DOES NOT take place in a
structured way.
HI children are discouraged from wearing
body-worn aids.
HI is NOT discussed formally with mainstream
children & they are NOT given information
about how to communicate with HI children
unless they request it.

Childrens attitudes toward disabled peers

341

Method
Design
The study employed a cross-sectional design involving the type of disability and method
of inclusion in each school, thus yielding a conceptual 3 6 2 between-subjects design:
Type of Disability (LD, HI, PD) 6 Method of Inclusion (`categorised , `decategorised ). Two additional schools which had no disabled children were used as controls.
The sample included two age groups (infant and junior) and equal numbers of girls and
boys. Because neither age nor sex of participating children is directly relevant to our
current concerns they are ignored in analyses reported below. For information,
however, no relevant age or sex differences were found.
Participants
A total of 256 children from two National Curriculum (NC) years one/two (infant aged
5 7 years, mean age 6.1 years) and three/four (junior aged 8 11 years, mean age 9.9
years), 128 girls and 128 boys. Ten children moved or were absent for at least one of the
three sessions. The children were randomly selected from within four classes in each of
eight schools in a large LEA in Southeast England. This sampling method was used to
reduce possible contaminating effects of idiosyncrasies of schools and class teachers.
Parental permission to participate in the study was obtained prior to random selection.
Procedure
Stimuli: The study utilised stimuli reported in past research (Maras & Brown, 1996).
Data described here used stimulus photographs of unknown children with (DIS) and
without (ND) disabilities to elicit children s reactions [standard stimuli (SS)]. The
photographs showed children with physical disabilities (PD) sitting in wheelchairs,
children with hearing impairment (HI) wearing body-worn hearing aids and children
with learning disabilities (LD) who had Down s Syndrome. Piloting determined that the
different disabilities were obvious to children of this age. The non-disabled children
were matched with the disabled children for age, sex and other features such as posture.
These standard stimuli photographs of unknown children were modified to show the
children in pairs and groups.
Photographs were also taken of all known non-disabled and disabled children (KC)
in the participating classes; these were utilised to determine sociometric choice and
preference.
Measures: Three measures used in this research were also utilised in our past work
(Maras & Brown, 1996). The first measure consisted of five `post-boxes into which the
children were asked to post the photographs of the known and unknown individual
children depending on how much they wanted to play with them (`always, `a lot ,
`sometimes , `not much or `never ), thus supplying a measure of sociometric preference.
The second, a five happy/sad faces Likert type scale, was modified for measuring affect
or liking. The third measure consisted of five different sized balloons. This was designed
for measuring amount of certain physical and psychological attributes.
Interview procedure: As in the previous study (Maras & Brown, 1996), participating
children used the measures and stimuli for two tasks described below. In addition, semistructured interviews were carried out with each child to elicit pre-existing stereotypes

342

Pam Maras and Rupert Brown

and attitudes about disability generally and then specifically about HI, LD and PD.
Prior to data collection the children were familiarised with the measures and the
researcher. Data collection was carried out over two sessions. All sessions were run in
the same way. Each child was seen individually in a quiet place away from the
classroom and verbatim instructions were used (Maras, 1993; Maras & Brown, 1996).
The children first responded to the measure of sociometric preference using both known
(KC) and unknown (SS) stimuli individually. They then responded to the evaluative
measures of psychological and physical attributes in respect of stimuli photographs of
the unknown children with and without disabilities (SS) which were presented in boy/
girl matched pairs (i.e., LD, HI, PD & ND). Order of presentation for the sociometric
preference task was random. For the evaluation task the disabled/non-disabled stimuli
were presented in a systematically ordered manner.
Results
These two procedures generated the various dependent measures for the study: the
amount of liking for and evaluation of abilities of each of the disability target categories
(LD, HI, PD, none) and sociometric preference for these categories and known peers.
Although the study was originally conceived of as a 3 6 2 (Type of Disability 6
Method of Inclusion) design, in order to incorporate the control schools, statistically it
was simpler to treat the eight participating schools as different levels of a single factor,
and then to include Disability of Stimuli as a within-subjects factor1 . Thus the initial
overall analyses adopted an 8 6 [4] mixed ANOVA strategy: School 6 [Stimuli]. Sex
and Age were included as between-subjects variables but are not reported below for
simplicitys sake. None of the results we report was qualified by these factors (i.e., there
were no reliable interactions involving sex and age).
We first present general findings from the study. Several of these analyses relate
directly to our theoretical perspectives on contact but are necessarily quite complex in
nature. In order to show the trends more clearly, we then focus exclusively on the data
from schools where children who are hearing impaired (HI) are included since it is in
these schools where the pattern of findings is most clear.2
Six mixed model ANOVAs (described above) were run on the intergroup attitude
indices (hearing ability, school work, physical education ability, running, hardworking,
thinking, liking, and `play with) using ratings of SS as the dependent measure. Given
that our hypotheses concern the effects of different kinds of inclusive practice adopted
by schools on intergroup attitudes, our primary interest is in statistical interactions
involving the School and targeted Stimuli variables. Such interactions indicate attitudes
toward particular stimulus groups vary as a function of the type of integrated contact
the participating children had experienced. Before discussing such interactions, it is
worth noting one consistent and highly significant finding which was observed on all
measures: without exception, a main within-subjects effect for Stimuli (ps < .001)
indicated that the three disabled groups were consistently evaluated as less able (and
less likeable) than the ND children. On several of the indices this main effect provided a
reassuring validation of our stimulus materials i.e., the children clearly understood
and indicated to us that, for example, ND children could `run better than children with
physical disabilities. Nevertheless, it was noticeable that such `realistic biases also

Childrens attitudes toward disabled peers

343

generalised inappropriately to other categories of disability e.g., children with LD and

HI were also judged to be able to run less well than ND children. It was also clear that
on the affective dimension `liking, disabled children were generally (and significantly)
less liked than non-disabled stimuli.
Of central interest in this paper are differences between rating of stimuli as function
of different contact experiences (i.e., different schools). Such results would demonstrate
a significant interaction involving type of school and stimuli. Such interactions were
observed on five dependent measures: `thinking (F (21,660) = 5.51, P < .001);
`hearing (F (21,657) = 3.95, p < .001); `running (F (21,690) = 3.10, p < .001);
`Physical Education (F (21,666) = 1.78, p < .05); and `liking (F (21,669) = 1.62,
p < .05); this last effect also involved sex of subject. To present the full results from all
these interactions would be both complex and burdensome for the reader. Thus we have
chosen to exemplify these effects by reporting fully the effects observed in the schools
where children who are HI are included. To interpret the significant interactions
reported below we employed a relatively conservative procedure. This involved the
assessment of pair-wise differences within each type of school (decategorised,
categorised and control) using Tukeys HSD post hoc procedure.
Results from the schools where hearing impaired children are included
Liking: As noted above, disabled children were rated as significantly less likeable than
ND children in all of the schools. As can be seen below (Figure 1) the difference
between ratings of disabled and ND children was greatest in the school where contact
was categorised and this difference is consistent for all three types of disability.
Hearing: When asked how well unknown children can hear (a dimension directly
relevant to the contact situation in these schools) all of the children indicated that the
HI children could hear significantly less well than the other disabled groups (LD & PD).
The children who had had contact rated the HI children lower than the control children
who had had no contact perhaps indicating a first effect of contact could be to make
the children more aware of the particular disability. Although the pattern is much the
same in the three types of schools, the differences are much sharper in the categorised
school than the decategorised and control schools. In all three conditions the children
attribute a degree of hearing impairment to the groups to which it is not relevant (LD &
PD); this is especially the case for PD stimuli children (Figure 2).
Schoolwork: Now we consider a dimension that is more relevant to the general school
context schoolwork. Here, again, there is a general trend to downgrade all of the
disabled children, particularly the PD group in the categorised and control schools (see
Figure 3). Contact, however, seems to have produced differing effects. In the control
schools the HI children are rated significantly higher than the PD children and there is
no significant difference between the children s ratings of the LD and PD children. In
the decategorised school there is no significant difference in the ratings of any of the
disabled groups. In the categorised school the pattern is somewhat different. As with
their rating of hearing, there was far more differentiation between disabilities and the
PD children are rated significantly lower than both the HI and LD children (Figure 3).
Thinking: Finally we consider a dimension that is more abstract and less directly
relevant to these particular contact situations thinking. Once again, all three disabled
groups are rated significantly lower than the non-disabled group. In the control schools

344

Pam Maras and Rupert Brown

Figure 1. Liking for SS in schools where children with hearing impairment are
integrated and control schools

LD = learning disability; HI = hearing impairment; PD = physical disability; ND = no

disability. Bars for each type of disability not sharing the same subscript in each type of contact
are significantly different by Tukeys HSD test, p < .05.

HI children are rated significantly higher than PD and LD between whom there is no
significant difference. A similar picture emerges in the decategorised school with no
significant difference between any of the disabled groups. In the categorised school an
entirely different pattern can be seen. Here, PD children and more importantly HI
children (a disabled group with whom they have contact), are rated significantly less
able to think than LD children (to whom the dimension is relevant) (Figure 4).
The data described so far have presented a somewhat negative picture of contact. The
most deleterious effects of contact appear to be in the categorised school in which the
children differentiate between the disabled groups more than in the decategorised and
control schools and downgrade the disabled group they know (HI) on dimensions that
are not relevant (schoolwork and thinking). In addition, in the categorised school, the
sharper differences spill over onto another disability (PD). Furthermore, children in all
of the contact situations clearly differentiate between disabled children generally and
non-disabled children.
Sociometric preference (`play with data): Data on the participating children s
sociometric preferences for known and unknown DIS and ND peers were then
analysed. These revealed a number of significant main effects and interactions but none
of these involved the simple School x Stimuli interaction of interest in this paper and so
will not be discussed further. Thus for sociometric preferences disability seemed not to
be a relevant factor in the children s judgments. Suffice it to say that the most powerful
results observed on these measures involved Sex x Gender of stimuli interactions,

Childrens attitudes toward disabled peers

345

Figure 2. Hearing rated for SS in schools where children with hearing impairment are
integrated and control schools

LD = learning disability; HI = hearing impairment; PD = physical disability; ND = no

disability. Bars for each type of disability not sharing the same subscript in each type of contact
are significantly different by Tukeys HSD test, p < .05.

reflecting the well documented finding of preference for same sex playmates amongst
young children (La Freniere, Stayer, & Gauthier, 1984; Maccoby & Jacklin, 1976). For
both known and unknown children these interactions (F (1,168) = 235.54 and
F(1,221) = 278.88, both p < .0001) completely masked all other statistical effects.
Attitude generalisation: It will be recalled that attitude generalisation from the known
contact person to other members of his or her category is the desired outcome of both
of the models of contact considered earlier. One index of this is provided by correlating
the participants ratings of how much they would like to play with known and unknown
same-sex children by gender. This strategy was employed to rule out confounding
effects of children s own gender preference referred to above. There was a significant
correlation between ratings of known and unknown girls in the categorised school. In
the schools with decategorised contact the correlations were lower and significantly
smaller than those in the categorised school. A similar pattern was observed amongst
the boys although the difference between the correlations just failed to achieve
conventional levels of statistical significance. Correlations were also carried out on the
children s mean ratings of the known and unknown non-disabled children. Here all but
one of the correlations were significantly different from zero at the .01 level and Z tests
revealed no significant differences between any of the correlations in the categorised
and decategorised schools. This indicates that the difference in generalisation of

346

Pam Maras and Rupert Brown

Figure 3. Schoolwork rated for SS in schools where children with hearing impairment
are integrated and control schools

LD = learning disability; HI = hearing impairment; PD = physical disability; ND = no

disability. Bars for each type of disability not sharing the same subscript in each type of contact
are significantly different by Tukeys HSD test p < .05.

attitudes visible in Table 2 was confined to the situationally relevant stimuli category
i.e., those with HI and was not a generic effect.

Discussion
In this paper we have reported findings from a large cross-sectional study involving
several schools practising different policies of inclusion for children with disabilities.
The focus was on mainstream children s attitudes towards their peers with disabilities
and the first notable result was that, in general, the participants showed marked
differentiation in their evaluations of, and liking for, the standard photographic stimuli
used to measure attitudes. As can be seen in Figures 1 4, this differentiation took the
form of a consistently more favourable attitude towards the ND stimuli than all three
disabled groups. Notice how the right hand bars in these figures are always significantly
higher than the remaining three bars. It is against this background of generic in-group
preference that the effects of different types of intra-school contact between children
with and without disabilities must be assessed.
These effects were indicated by the presence of statistically reliable interactions
between school and stimuli on most measures. The simplest way to describe these
interactions is to note that, in schools in which there was a concentration of children
with a particular form of disability, those that down-played the salience of that
disability category what we have termed decategorised schools tended to have

Childrens attitudes toward disabled peers

347

Figure 4. Thinking rated for SS in schools where children with hearing impairment are
integrated and control schools

LD = learning disability; HI = hearing impairment; PD = physical disability; ND = no

disability. Bars for each type of disability not sharing the same subscript in each type of contact
are significantly different by Tukeys HSD test, p < .05.

children with less differentiated (and less biased) attitudes than those who placed more
emphasis on disability (categorised schools). In fact, the decategorised schools generally
tended to resemble the `no contact control schools in the profile of their intergroup
attitudes on different evaluative dimensions. This pattern is well exemplified in the
schools with significant numbers of children with hearing impairments, whose results
we have presented in detail (Figure 1 4). Note how the differences between the HI and
ND bars are always greater in the categorised schools than in the decategorised or
control schools.
At first glance, such findings clearly offer more support to Brewer and Miller s (1984)
contact model than to Hewstone and Brown (1986). As predicted by Brewer and Miller,
where a social category is psychologically less significant it loses its power to organise
(and bias) people s attitudes. Nevertheless, despite this empirical support, from an
applied perspective the Brewer-Miller model does not offer a very optimistic prognosis
for school inclusion policies. For, although children in the decategorised schools had
less biased attitudes than those in the categorised schools, these attitudes were not on
the whole any more favourable than those shown by children in the control schools.
Against this baseline, then, even decategorised contact seems to be having scant effect
on mainstream children s attitudes toward disability.
Inclusion policies more closely resembling Hewstone and Brown s (1986) version of
the contact hypothesis would appear to have even lower chances of success, if these

348

Pam Maras and Rupert Brown

Table 2. Correlations between children s ratings of `play with known and unknown
children who are hearing impaired
Contact with children who are
Hearing Impaired
Categorised
De-categorised
Difference between schools (Z
score) N = 64

Correlation between known and unknown stimuli

photographs of children who are Hearing Impaired
Girls
Boys
.56**
.00
2.41**

.32**
7.06
1.52

*p < .05; **p < .01; ***p < .001

findings are a guide. A plausible reason for the more sharply differentiated and negative
intergroup attitudes observed in the categorised schools is that the contact which
occurred here was hardly optimal in terms of Allport s (1954) criteria. Qualitative data
from the interviews indicated that the children were not given much information about
the nature of different disabilities and many felt uncomfortable about meeting their
disabled peers. In other contexts, anxiety has been shown to be negatively associated
with favourable intergroup attitudes (Greenland & Brown, 1999; Islam & Hewstone,
1993; Stephan & Stephan, 1985), and it would be surprising if similar processes were
also not operative in school contexts. Moreover, by common consent, class sizes in
these schools were large, thus creating demands on teachers time and precluding the
possibility of frequent and effective co-operative learning activities. There was a striking
contrast between the generally negative effects of contact in these schools and the
positive outcomes observed in an earlier study of a contact situation involving regular
structured co-operative encounters between children with and without severe learning
difficulties (Maras & Brown, 1996).
Despite the practical difficulties of trying to apply Hewstone and Browns (1986)
model in sub-optimal conditions, theoretically their hypothesis received at least partial
support from the observation that attitude generalisation appeared more in evidence in
the categorised than the decategorised schools. This was shown by the stronger
correlations between sociometric preferences for known and unknown peers with
disabilities in the former than the latter schools. A likely reason for this is that the
heightened salience of the disability categories in categorised schools permitted
psychological transfer of feelings about a known exemplar category to other members
of that category not yet met. Of course, in this instance the generalisation was probably
of negative attitudes, but this need not always be the case (Maras & Brown, 1996; Van
Oudenhouven, Groenewoud, & Hewstone, 1996; Brown et al., 1999).
Before concluding it is appropriate to enter four caveats about our research design
which qualify our conclusions. Because the study was conducted in a real-life context,
exploiting ongoing policy differences in several schools, we do not have the luxury of a
true experimental design in which children were randomly assigned to different contact
experiences. In addition, our characterisation of the different inclusion policies as
involving categorised and decategorised contact was undoubtedly an imperfect one

Childrens attitudes toward disabled peers

349

since there were probably other variables confounded with this distinction. Still, the
consistency and (statistical) strength of our findings across several different measures
give us some confidence that they are not completely spurious and can be meaningfully
related to the theoretical and policy issues we identified at the outset. Our main
conclusion is that most schools we studied were not providing the conditions which
promoted the most effective forms of contact between different groups of their students.
The challenge for teachers and education policy-makers alike is to re-create in
mainstream classrooms the kinds of co-operative and equal status learning contexts
within a national curriculum that currently allows little time for alternative methods.
In conclusion, we have intentionally not in this paper linked our theoretical
conceptualisations to pedagogy. However, given current increased interest in links
between research, theory and pedagogy it is interesting to note that current debates on
`unique vs. `generalisable differences as recently summarised by Lewis and Norwich
(1999) bear more than a passing conceptual resemblance to the social psychological
perspectives outlined in this paper. Perhaps, therefore, work aimed at linking delivery
and practice to work on social psychological processes might prove a useful way
forward. The challenge of course is accommodating moral often posing as theoretical
tensions between views that purport inclusion per se and those that see education (and
learning) for all as a main goal for the future.
NOTES
1

For illustrative purposes scores for the control schools are derived from the mean scores of both
control schools in Figures 1 4. In addition a mean score was also computed for the three nondisabled pairs of stimuli. Prior analysis revealed that this strategy was appropriate, as there was
little difference between the childrens evaluations of the ND pairs when they were included in the
analysis separately.
2
Results from the other schools show broadly similar patterns and can be obtained from the first
author.

Acknowledgment
The research reported here was funded by grant (#F2368) from the Leverhulme Trust.

References
Acton, H.M., & Zarbatany, L. (1988) Interaction and performance within cooperative groups:
Effects on nonhandicapped students attitudes toward their mildly mentally retarded peers.
American Journal of Mental Retardation, 95, 16 23
Allport, G.W. (1954 ed.) The nature of prejudice. Cambridge/Reading, MA: Addison-Wesley.
Armstrong, B., Johnson, D.W., & Balow, B. (1981). Effects of cooperative vs. individualistic
learning experiences on interpersonal attraction between learning-disabled and normalprogress elementary school students. Contemporary Educational Psychology, 6, 102 109.
Barnes, C. (1996). Disability and the myth of the independent researcher. Disability & Society, 11,
107 110.
Bettencourt, B.A., Charlton, K., & Kernahan, C. (1997). Numerical representation of groups in
cooperative settings: Social orientation effects on ingroup bias. Journal of Experimental Social
Psychology, 33, 630 359.
Brewer, M.B., & Miller, N. (1984). Beyond the contact hypothesis: Theoretical perspectives on

350

Pam Maras and Rupert Brown

desegregation. In N. Miller & M.B. Brewer (Eds.), Groups in contact: The psychology of
desegregation (pp. 281 302). New York: Academic Press.
Brinker, R.P. (1985). Interactions between severely mentally retarded students and other students
in integrated and segregated public school settings. American Association on Mental Deficiency,
89, 6587 6594.
Brown, R., Vivian, J., & Hewstone, M (1999). Changing attitudes through intergroup contact:
The effects of membership salience. European Journal of Social Psychology, 29 741 764.
Chesler, M.A. (1965). Ethnocentrism and attitudes toward the physically disabled. Journal of
Personality and Social Psychology, 2, 877 882.
Department for Education (1993). Education Act. London: HMSO.
Department for Education (1994). Code of practice on identification and assessment of special
educational needs. London: HMSO.
Department for Education and Employment (1998). Meeting special educational needs: A
programme of action. London: HMSO.
Department for Education and Employment (1997). Excellence for all children: Meeting special
educational needs (Green Paper). London: HMSO.
Department of Education and Science (1978). Report of the committee of enquiry into the
education of handicapped children and young people. (Warnock Report). London: HMSO.
Department of Education and Science (1981). Education Act. London: HMSO.
Dockrell, J., Logotheti, A.E., & Magiati, I. (1998). Primary school childrens representations of
special educational needs: Patterns of awareness and social/locational influences in Greek
children. British Psychological Society Annual Conference of the Education Section, Exeter.
Furnham, A., & Pendred, J. (1984). Attitudes towards the mentally and physically disabled.
British Journal of Medical Psychology, 56, 179 187.
Greenland, K., & Brown, R. (1999). Categorisation and intergroup anxiety in contact between
British and Japanese nationals. European Journal of Social Psychology 29, 503 521.
Hewstone, M., & Brown, R. (1986). Contact is not enough: An inter-group perspective on the
Contact Hypothesis. In M. Hewstone & R. Brown (Eds.), Contact and conflict in inter-group
encounters (pp.1 44). Oxford: Blackwell.
Islam, M.R., & Hewstone, M. (1993). Dimensions of contact as predictors of intergroup anxiety,
perceived ingroup variability and outgroup attitudes: An integrative model. Personality and
Social Psychology Bulletin, 19, 700 710.
Johnston, M. (1997). Representations of disability. In K.J. Petrie & J.A. Weinman (Eds.),
Perceptions of health and illness: Current research and applications (pp. 189 212). Singapore:
Harwood Academic Publishers.
Johnson, D.W., & Johnson, R.T. (1981). The integration of the handicapped into the regular
classroom. Effects of cooperative and individualistic instruction. Contemporary Educational
Psychology, 6, 344 353.
Johnson, R. et al. (1979). Interaction between handicapped and non handicapped teenagers as a
function of situational goal structuring: Implications for mainstreaming. American Educational
Research Journal, 16, 161 167.
La Freniere, P., Stayer, F.F., & Gauthier, R. (1984). The emergence of same sex affiliative
preferences among school peers: A developmental/ethological perspective. Child Development,
55 1958 1965.
Lewis, A., & Lewis, V. (1987). The attitudes of young children towards peers with severe learning
difficulties. British Journal of Developmental Psychology, 5, 287 292.
Lewis, A., & Norwich, B. (1999). Mapping a pedagogy for special educational needs. Research
Intelligence, 69, 9 13.
Lunt, I., & Norwich, B. (1999). Can effective schools be inclusive schools? London: Institute of
Education, London University.
Maccoby, E.E., & Jacklin, C.N. (1976). Gender segregation in childhood. Advances in Child
Development and Behaviour, 20, 239 287.
Maras, P. (1993). The integration of children with disabilities into the mainstream: Effects of school
and age on mainstream childrens attitudes toward disability. Unpublished doctoral thesis,
University of Kent.

Childrens attitudes toward disabled peers

351

Maras, P. (1996). `Id rather have dyslexia: Perceptions of EBDs. Educational and Child
Psychology, 13 (1), 32 43.
Maras, P., & Brown, R. (1996). Effects of contact on childrens attitudes towards disability: A
longitudinal study. Journal of Applied Social Psychology, 26, 2113 2134.
McConkey, R., McCormick, B., & Naughton M. (1983). A national survey of young peoples
perceptions of mental handicap. Journal of Mental Deficiency Research, 27, 171 183.
Miller N., Brewer, M.B., & Edwards K. (1985). Co-operative interaction in desegregated settings:
A laboratory analogue. Journal of Social Issues, 41, 63 79.
Norwich, B. (1990). Special needs in ordinary schools. Reappraising special needs education.
London: Cassell.
Norwich, B. (1994). Segregation and inclusion. English LEA statistics 1998 1992. Bristol: Centre
for Studies in Integration in Education.
Norwich, B. (1997). A trend towards inclusion. Bristol: Centre for Studies in Integration in
Education.
Oliver, M. (1990). The politics of disablement. Basingstoke: Macmillan.
Shakespeare, T., & Watson N. (1997). Defending the social model. Disability and Society, 12 (2),
293 300.
Siperstein, G.N., & Chatillon A.C. (1982). Importance of perceived similarity in improving
childrens attitudes towards mentally retarded peers. American Journal of Mental Deficiency, 86
(5), 453 458.
Spillers, C.S. (1982). An investigation of childrens attitudes towards physically disabled peers.
Mid-American Review of Sociology, 7 (1), 55 69.
Stephan, W., & Stephan, C. (1985). Intergroup anxiety. Journal of Social Issues, 41, 157 175.
Strohmer, D.C., Grand, S.A., & Purcell M.J. (1984). Attitudes towards persons with a disability:
An examination of demographic factors, social context, and specific disability. Rehabilitation
Psychology, 29 (3), 131 145.
Van Oudenhouven, J.P., Groenewoud, J.T., & Hewstone, M. (1996). Cooperation, ethnic salience
and generalisation of interethnic attitudes. European Journal of Social Psychology, 26, 649 662.
Vivian, J., Hewstone, M., & Brown, R. (1997). Intergroup contact: Theoretical and empirical
developments. In R. Ben-Ari & Y. Rich (Eds.), Understanding and enhancing education for
diverse students. An international perspective (pp. 13 46). Jerusalem. Bar-I1am University Press.
White, J., & Barber, M. (1997). Perspectives on school effectiveness and school improvement.
London: Bedford Way Papers.
Received 3 September 1999; final version received 20 January 2000