Because the study of the self-concept is, in essence, the study of individual differences it follows that it can be undertaken most readily in situations where such differences are maximised. One way of doing this is by considering the impact of disability upon the person's concept of self. By studying children who have differing kinds and levels of disability and by comparing their concepts of self with those who do not have such an impairment, the impact of individual differences upon notions of self can be more effectively analysed. For example, children with learning disabilities might be expected to have a very different view and understanding of their cognitive skills and abilities than those who are developing typically. How these different capabilities translate to a concept of self and notions of self-worth can be seen to have significance not only for model and theory building in the general field of self-concept but might also be expected to have practical implications for interventions to raise self-esteem.

Central to any consideration of the constructs of self is how the four terms - self-concept, self-image, self-esteem and self-worth, often used interchangeably in the literature, relate one to the other and how they should be delineated. Indeed, distinguishing among them has proved difficult on both a theoretical and an empirical level. By studying the impact of the specific functional loss which results from a disability upon the individual's construct of self, a clearer picture of the mechanisms involved may be obtained. Thus, for example, children, responding to a self-concept scale such as the Self Description Questionnaire (developed by Marsh and Smith in 1981) are asked, amongst other items, to state whether they are good at reading and whether they look forward to it. Analysis of the responses of children who have a learning disability will help to determine the extent to which the individual's image of his/her skills is based upon functional ability rather than upon an idealised view of what he or she is capable of achieving. The role of an ideal in determining our image of self is one which Harter (1985) has explored extensively, particularly with respect to the young child, and which can be further elucidated in this way.

If a developmental perspective is taken, additional insights can be derived from studying the self-concept of children with disabilities. Thus, it is probable, as Silon and Harter (1985) suggest, that children with intellectual disabilities have less clearly defined notions of self and self-image than either typically developing children or children with physical, learning or sensory disabilities but no intellectual deficit of the same chronological age. For them, delay in the formation of a stable and complex self-concept is as likely as it is in the academic areas of reading or mathematics.

Finally, through this approach the processes involved in determining levels of self-image, self-esteem and feelings of self-worth can also be further explored. In our society, physical appearance and motor skills are highly valued and are important factors in determining the popularity of children. Children with physical disabilities, if they accept these values, may be condemning themselves to low self-esteem. The same would also be true for children whose disability has typical and distinctive characteristics as is the case with, for example, Down syndrome. Research has not, however, consistently found that individuals with disabilities have lower levels of self-esteem and self-worth than their typically developing peers (King et al., 1993; Chapman, 1988; Silverman & Zigman, 1983).

What has been found in a number of studies, however, is that those with disabilities score lower on measures of self directly affected by their disability (Chapman, 1988; Cooley & Ayres, 1988; Grolnick & Ryan, 1990; King et al., 1993). More specifically, Grolnick and Ryan (1990) found that students with learning disabilities scored lower on measures of cognitive competence and academic self regulation relative to the non-disabled control groups but that they did not differ on general self perceptions of control or competence.

Similarly, King et al. (1993) in their study of adolescents with physical disabilities found that females were lower in perceived social acceptance, athletic competence and romantic appeal than the normative sample whilst males were lower in perceived scholastic competence, athletic competence and romantic appeal. As already noted, they did not differ from the normative sample with respect to global self-worth. The distinction between self-image and self-worth is made clearly. These adolescents were realistic about their skills and capabilities in the areas directly affected by their physical disability and yet were able to maintain their feelings of self-worth. Here again, then, the study of the impact of disability upon the child's perceptions allows a clearer understanding of the factors influencing the development of self-concept.

It is therefore argued that comparing self-concept development in those who have a disability with those who do not provides a powerful way of clarifying the underlying constructs. It is, consequently, of some surprise that such an approach has not been more widely used. For the most part, where disability has been a factor in research on self-concept it has been simply another dependent variable used to distinguish those with disabilities from their non-disabled peers. The exception to this has been the work of Harter and her colleagues (Silon & Harter, 1985; Renick, 1985; Mayberry, 1989; Renick & Harter, 1989) who have seen studying the impact of disability as a means of clarifying our knowledge of the structure of the self.

The data presented here derive from a larger study which has aimed to extend this approach by exploring and comparing the development of self-concept in children with learning, physical and intellectual disabilities with that of their typically developing peers.

Specifically, two questions were considered for this analysis:

1. What is the impact of disability on children’s views of their levels of competence and social acceptance? And
2. Do children with disabilities have lower self-worth than their typically developing peers?

The sample of children aged six to twelve years was drawn from six Sydney Metropolitan schools with additional children with cerebral palsy identified through the voluntary organisation from which they were receiving services. Details of this sample of 247 children are set out in Table 1.

The majority of the children tested were in Grade 3. The uneven distribution of the sample across the primary years was a function both of the classes the schools made available for testing and parents’ willingness to have their children participate. This was particularly so for the children with physical disabilities, with the overwhelming majority of children entering the study through this route being aged eight to ten years. Interestingly, this has been suggested to be a critical time for children with disabilities (Dunn et al., 1988, p.249). Harter (1990a) has also pointed to age eight as the time when the child begins to formulate a more complex sense of self. The proportion of males to females within the disability groups mirrors the greater number of boys with disabilities within the general population.

The characteristics of the groups of children were:

Typically developing children (n = 173): All the children in this group were attending local, community schools. None had been identified as having a developmental disability or delay.

Children with physical disabilities (n = 23): All the children had been identified as having cerebral palsy. Of these, six used a wheelchair; one was attending a special school and five a special unit within a regular school with some integration into the regular classroom. All had at least some language with two communicating primarily through symbol boards.

Children with intellectual disabilities (n = 25): All the children in this group had been identified as having a mild to moderate level of intellectual disability. All were enrolled in special units within regular schools with most spending some time in the regular classroom.

Children with learning disabilities (n = 26): All the children in this group had been identified by their schools as experiencing specific difficulties particularly with respect to learning reading. They had also been diagnosed, initially by a paediatrician, as having ADD/ADHD. All were included in the regular classroom.

The Self Description Questionnaire (SDQ) was developed by Marsh and Smith (1981) and has become perhaps the most frequently used scale in studies concerned with self-concept in older children, adolescents and adults. The SDQ is based on the theoretical model of self-concept developed initially by Shavelson et al. (1976). Of prime importance to the present study is the fact that the SDQ takes account of the multi-dimensionality of the individual's view of self allowing comparisons within and between subjects with respect to different facets of development and competence. As Kistner, Haskett, White and Robbins (1987, p.38) have argued, such specificity is needed when considering the impact of disability upon the development of self-concept.

There are 76 items with children being asked to respond to statements about their skills and views of self along a five point Likert scale. The SDQ has sound psychometric properties which have been set out by Marsh (1987). Its wide use has meant that comparisons can be made against normed data. Although the length of the scale was of some concern in attempting to apply it with children with intellectual disabilities, it was seen as the most appropriate measure to look at the multi-dimensional nature of the self-concept.

The SDQ is normally group administered but can also be given individually. The children with intellectual disabilities were tested individually at their schools to ensure that they understood the questions and were able to follow the instructions. The children with physical disabilities were also tested individually with the exception of four children who were fully integrated into the regular classroom and thus tested with their classmates.

1. What is the impact of disability on children’s views of their levels of competence and social acceptance?

Internal consistency coefficients were calculated for each of the SDQ subscales using Cronbach’s alpha. These were found to be similar to those obtained by Marsh (1987) for the sample of typically developing children. Children with disabilities were not included in Marsh’s analysis. The results are given in Table 2 with the results obtained by Marsh (1987) set out in boldface type.

Of interest are the differences among the groups in the reliability coefficients obtained. Combining the disability groups serves to hide the differences which exist among them as a result of the differential impact of the disabilities themselves.

It should be noted that the internal consistency coefficients are considerably lower on the Parent Relations and General subscales for all the disability groups but especially for the children with intellectual disabilities on Parent Relations. Lower coefficients are also evident for the children with intellectual or learning disabilities on the Reading, Maths and School subscales. It is to be wondered whether their awareness of the particular impact of their disabilities is causing them to make very specific judgements and assessments of their levels of competence in the domains most affected. This view would tend to be borne out by similar coefficients being found for the other areas as were found for their typically developing peers.

Planned comparisons were carried out on the data obtained from the four groups to test specific hypotheses related to the children’s views of their competence and acceptance. The rationale underlying the predictions made is that where disability directly affects functioning in a domain, children’s perceived competence and acceptance will be lower than that measured for their typically developing peers.

The Levene Test for Homogeneity of Variance was carried out for each of the subscales before the analyses were conducted. Significant results were found for the following subscales: Parental Acceptance, Maths Ability, Reading Ability and School Ability. The assumption of equality of variance was therefore met for three of the seven subscales obtained on the SDQ. The results for the remaining four subscales should be treated with some caution. It should be noted, however, that analyses were also conducted using the Jonckheere-Terpstra test for k independent samples. The results obtained were very similar to those obtained using planned comparisons. The robustness of techniques involving the analysis of variance would thus seem to be supported.

1. Perceptions of Physical Abilities

It was predicted that the group of typically developing children would have higher perceived physical abilities than their peers with physical, intellectual or learning disabilities since physical skills were seen as being affected by all three types of disability being studied.

The results of the planned comparisons are set out in Table 3.

The results obtained do not support the hypotheses; disability is not seen as affecting this area of the children’s self-image. Given the impact which their disabilities are having on this skill area, this is an interesting finding and is considered further below.

2. Perceptions of Physical Appearance

Because the nature of physical and intellectual disabilities is such that they are likely to result in visible differences, it was hypothesised that children in these two groups would have lower perceived attractiveness than their typically developing peers. No difference was predicted between the scores of those with learning disabilities and those of their typically developing peers.

Table 4 sets out the results of the planned comparisons. One of the hypotheses was supported with the other two comparisons yielding significant results in the opposite direction to that predicted.

That the two groups of children with physical or intellectual disabilities perceive themselves as more attractive is of interest. It is worth noting, however, that the mean obtained for Physical Appearance by the typical group is lower than any of their other subscale scores. In this instance the differences found are due both to the typically developing group’s having their lowest score in this domain and those with either physical or intellectual disabilities seeing themselves more positively in this area of self. The mean for those with learning disabilities is similar to the typical group.

From anecdotal evidence obtained from the comments of the children with intellectual or physical disabilities as they completed the relevant items, it seems likely that their parents and teachers have emphasised their attractiveness and worth in this respect. As a result, it may be that they are less susceptible to peer values as they apply to physical attractiveness.

3. Perceptions of Peer Relations

Disability is seen as affecting not only the individual’s ability to engage with peers and form friendships but also other children’s reactions to them. It was, therefore, predicted that typically developing children would have higher perceived peer acceptance than their peers with physical, intellectual or learning disabilities.

The results of the planned comparisons are set out in Table 5.

None of the hypotheses was supported in this area. The predicted difference resulting from perceived lack of acceptance and poorer social skills was not found; the children with disabilities held similar views of their level of peer acceptance to their typically developing peers. If these were found to translate to their interactions with peers the implications for successful inclusion would be very positive.

One other factor does need to be considered in interpreting these findings; the children’s school placement. Since a number of the children with intellectual or physical disabilities were in special units, it may be that they are not looking to the wider school community in judging their level of acceptance. The reference group they are using may well be critical.

4. Perceptions of Parent Relations

The children’s disabilities are not predicted to have a differential effect on their relationship with their parents. As a consequence, no differences were predicted between the groups of children with disabilities and their typically developing peers with respect to their perceived levels of parental acceptance.

The results yielded are set out in Table 6. All three hypotheses were supported.

The results are in accordance with prediction. It should be noted, however, that this is one of the areas for which homogeneity of variance was not supported. The greatest variance, in this instance, was associated with the largest sample (typically developing children); a result which would be taken to result in the F test being too conservative. Some caution is therefore indicated in interpreting the results.

5. Perceptions of Maths, Reading and General School Ability

The analyses carried out for the Maths, Reading and General School Ability subscales have been set out together both to simplify presentation and because the predictions made with respect to the three areas share a common rationale. The General School Ability subscale comprises a set of separate items related to the children’s perceptions of their competence within the school curriculum generally.

It was predicted that, since their disabilities would have a direct effect on their functioning in these areas, children with either intellectual or learning disabilities would score themselves lower than would their typically developing peers on Maths, Reading and General School ability. No differences between the scores of the children with physical disabilities and their typically developing peers on Maths, Reading and General School Ability were predicted.

The data obtained from these analyses are set out in Table 7 below.

With respect to the children with physical disabilities, the hypotheses were supported for Maths and General School Ability but not for Reading Ability where they saw themselves as performing less well than did their typically developing peers. Homogeneity of variance was not met for this last comparison with the larger variance associated with the sample of children with physical disabilities. This would suggest that the test might be too liberal. When the separate variance estimate is used for that comparison a p value of .066 is yielded. Whether the difference found is meaningful is, therefore, questionable.

None of the hypotheses were supported for the group of children with intellectual disabilities. The failure to find a significant difference between the children who are typically developing and those with intellectual disabilities is of interest. Despite the direct impact which their disability is having on their learning, these children still perceive themselves as succeeding in the school environment. Their segregation in special units and the resultant differences in their curricula may be factors in their judgements of their levels of competence. It is worth emphasising that it is irrelevant whether these children are succeeding academically or not by objective measures, what is important is their perception of how they are performing as it is this which will affect their feelings of self-esteem and self-worth. In this case the notion of the "looking glass self" (Shrauger & Schoeneman, 1979) and Marsh’s (1984) concept of the "big fish, little pond effect" may be acting very much to the children’s advantage as they gauge their level of skill.

The hypotheses relating to Maths and General School Ability were supported for the children with learning disabilities whilst the comparison related to Reading Ability showed no differences between their perceptions and those of their typically developing peers. It should be noted that, with respect to the children’s perceptions of their Maths Ability, homogeneity of variance was not met with the smaller variance associated with the typically developing group. This would be expected to result in an interpretation of the data which was too liberal. However, when the separate variance estimate is used for the comparison between the typically developing group and the children with learning disabilities, the difference remains statistically significant (p < .035). The result can, therefore, be accepted albeit with some caution.

Similarly, homogeneity of variance was not met for the General School Ability subscale. However, the significant difference found for the children with learning disabilities may be accepted with some caution since the larger variance was associated with the smaller sample. When the separate variance estimate is used for that comparison a p value of .015 is obtained.

Why the prediction that children with learning disabilities would see themselves as not performing as well in Reading as would their typically developing peers was not supported is puzzling. It may be that, as with the perceptions of the children with intellectual disabilities with respect to their perceptions of their maths and reading competence, they are using a different reference system.

What is notable, in considering the results for each of the subscales of the SDQ, is that the view that children have lowered perceptions of their competence and social acceptance in the areas affected by their disability has not been consistently supported. Indeed, the only area where the hypotheses were fully supported was Parent Relations. In the other areas, either no differences were found when they predicted or differences were found but not in the direction predicted (this was the case in the area of Physical Appearance where the typically developing children scored themselves lower than did their peers with Physical or Intellectual Disabilities). A number of potential explanations need to be considered. These include the possibility that the children with disabilities are simply unrealistic or idealistic in their assessment of their level of competence and acceptance or that they using a different reference system to determine their level of accomplishment. These notions warrant further study.

The General scale of the SDQ is derived from Rosenberg’s Self-Esteem Scale and is considered separately from the other components of the SDQ (Marsh, 1987). It purports to be a measure of the individual’s global sense of self-worth.

It was predicted that no differences would be found in the levels of General Self-Worth between the children with disabilities and their typically developing peers. What is being argued here is that children with disabilities are as likely as those who are developing typically to employ, or not, the processes of discounting and appropriate reference systems to maintain and enhance feeling of self-worth. The results of these comparisons are set out in Table 8.

The predictions were, therefore, fully supported. The lack of differences found among the groups in their estimations of their self-worth provides support for the view that the presence of a disability does not in itself result in lowered self-worth.

A number of conclusions can be drawn from the data as presented. Firstly, and importantly for future research using disability as a variable, different disability types should not be combined. To do so masks the specific impact of the disability upon the individual’s functioning and views of self. Differences which exist in reality may, also, not emerge.

Secondly, there is additional support for the view that disability does not in and of itself result in lowered self-worth. This is a critical consideration not only for those working in the disability field but also for any attempts to intervene to raise self-esteem.

What then needs to be determined is how individuals with clear deficits in their functioning manage to achieve positive self-worth. That the comparisons using the subscales of the SDQ did not consistently show the children with disabilities to have lower perceptions of their skills than their typically developing peers needs to be considered in this regard since it would be expected that they would recognise and react to the impact of their disability. It may be, of course, that the children are unrealistic in their evaluations, that they are subscribing to an idealised view of self. This will no doubt be true for a number of individuals whether they have a disability or not. Another explanation may, however, be possible.

As already stated, the data presented here are part of a much larger study. As a consequence, only part of the picture is being seen. The data can stand alone and, indeed, similar approaches abound in the literature on self-concept particularly in studies comparing different groups of individuals. However, what this approach does not do is to allow the children’s responses to be fully understood. It is argued that the analyses set out here should not be considered in isolation. Account must be taken of the reference system which the children are using as well as the other techniques by which they determine their views of self. Performance is just one factor in what is a complex process of self-evaluation. This view is reinforced by the fact that the young children from the preschool sample in the full study did show the impact of disability upon their perceptions of their skills and abilities (Johnston, unpublished). How these findings can be reconciled must therefore be addressed.

If the underlying process in evaluating the self can be understood, intervention to raise self-esteem can be predicated on a sound base.

Chapman, J.W. (1988) Learning disabled children's self-concepts, Review of Educational Research, 58, 3, 347-371.

Cooley, E.J. & Ayres, R.R. (1988) Self-concept and success-failure attributions of non-handicapped students and students with learning disabilities, Journal of Learning Disabilities, 21, 3, 174-178.

Dunn, N.L., McCartan, K.W. & Fuqua, R.W. (1988) Young children with orthopedic handicaps: Self-knowledge about their disability, Exceptional Children, 55, 3, 249-252.

Grolnick, W.S. & Ryan, R.M. (1990) Self-perceptions, motivation, and adjustment in children with learning disabilities: A multiple group comparison study, Journal of Learning Disabilities, 23, 3, 177-184.

Harter, S. (1985) Competence as a Dimension of Self-Evaluation: Toward a Comprehensive Model of Self-Worth in Leahy, R.L. (ed), The Development of Self, Orlando: Academic Press Inc.

Harter, S. (1990) Causes, Correlates and the Functional Role of Global Self-Worth: A Life-Span Perspective in Sternberg, R.J. & Kolligan, J. (eds), Competence Considered, New Haven: Yale University Press

King, G.A., Shultz, I.Z., Steel, K., Gilpin, M. & Cathers, T. (1993) Self-evaluation and self-concept of adolescents with physical disabilities, American Journal of Occupational Therapy, 47, 2, 132-140.

Kistner, J., Haskett, M., White, K. & Robbins, F. (1987) Perceived competence and self-worth of LD and normally achieving students, Learning Disability Quarterly, 10, 1, 37-44.

Marsh, H.W. (1987) The Self Description Questionnaire (SDQ): A Theoretical and Empirical Basis for the Measurement of Multiple Dimensions of Preadolescent Self-Concept: A Test Manual and a Research Monograph, unpublished manuscript.

Marsh, H.W. & Smith, I.D. (1981) Self Description Questionnaire, Sydney: University of Sydney.

Mayberry, W. (1989) Self-esteem in children: Considerations for measurement and intervention, American Journal of Occupational Therapy, 44, 8, 729-734.

Renick, M.J. (1985) The Development of Learning Disabled Children’s Self-Perceptions, Master’s Thesis, University of Denver, Colorado.

Renick, M.J. & Harter, S. (1989) impact of social comparisons on the developing self-perceptions of learning disabled students, Journal of Educational Psychology, 81, 4, 631-638.

Silon, E. & Harter, S. (1985) Assessment of perceived competence, motivational orientation and anxiety in mainstreamed and self-contained educable mentally retarded children, Journal of Educational Psychology, 77, 2, 217-230.

Silverman, R. & Zigmond, N. (1983) Self-concept in learning disabled adolescents, Journal of Learning Disabilities, 16, 8, 478-482.