Anne McMaugh and Ray Debus, University of Sydney.

Paper presented at the annual meeting of the Australian Association of Research in Education, Adelaide, December, 1998

While research and theory on the adjustment of children with chronic conditions and disabilities has grown considerably in recent decades, the study of peer relations within this population has been relatively neglected (La Greca, 1990). Consequently, our understanding of social adjustment within this population remains limited and equivocal. Similarly, our understanding of related social cognitive processes is limited. This paper examines the potentially important role of the self-efficacy construct in providing some insight into the role of cognitive processes in social adjustment. The development of a self-efficacy instrument, the Self Efficacy for Peer Interactions Scale (SEPI) is examined with preliminary pilot and research data.

Although potentially important for understanding children's peer relationships, the construct of self-efficacy has received very little attention in literature concerning children's social adjustment. Although the research on which this paper is based is concerned primarily with children with certain forms of chronic conditions†, it is important to note that social self-efficacy has been little explored in either this population or among children without such conditions. This paper aims to provide an overview of research on the adjustment processes of children with chronic physical conditions. The paper advances a rationale for the study of social self-efficacy and the paper reports the development of a social self-efficacy instrument, designed for use in a population of students with chronic physical conditions.

Although actual peer relationships have been less studied, the study of adjustment among children with chronic conditions has for some time been of interest to researchers. Researchers in areas of pediatric psychology in particular have taken great interest in how the process of living with a chronic condition or disability can affect childhood adjustment processes. In recent years several reviews (Eiser, 1990; LaGreca, 1990; Spirito, DeLawyer & Stark, 1991) and a significant meta-analysis (Lavigne & Fair-Routman,1992) have converged in pointing out that children with chronic conditions are at risk of experiencing maladjustment problems.

The area of peer relationships has been particularly acknowledged as a potential area of difficulty for children with chronic conditions (La Greca, 1990; Spirito, DeLawyer and Stark, 1991). While it is too early to conclude unequivocally that children with such conditions are generally at risk for social adjustment problems, it is suggested that certain problems that arise along with having a condition or disability may affect peer relationships. In particular Spirito, DeLaywer and Stark (1991) point out that conditions that affect the physical appearance or restrict activities may be most likely to impact upon relationships. Similarly, La Greca (1990) suggested that restrictions to physical activity, interruption of daily activities, altered or unusual physical appearance and forced lifestyle changes might influence the peer relationships of children with chronic conditions.

However, marked methodological deficiencies within the field of social adjustment studies have also been highlighted. Common criticisms have included the failure of research studies to examine social adjustment specifically (Spirito, DeLaywer and Stark, 1991; Nassau and Drotar, 1995; Thompson and Gustafson, 1996). Often claims about social adjustment have been made as an aside to a general study of psychosocial adjustment. Multiscale diagnostic tools such as the Child Behaviour Checklist [CBCL] (Achenbach & Edelbrock, 1984), which contains a small subscale of items regarding social competence, have been used widely to report social maladjustment among this population of children (Thompson and Gustafson, 1996). Perrin, Stein and Drotar (1991) point out that this instrument may not adequately conceptualise or represent social adjustment due to the small subscale of items and the tendency of the items to assess actual functioning rather than competence. This tendency to assess social functioning rather than competence has also led to the suggestion that this form of assessment of social adjustment may be inappropriate for children with disabilities (Perrin. Stein and Drotar, 1991; Thompson & Gustafson, 1996). Items on the scale that reflect school attendance, participation in physical activities and participation in household chores may be inappropriate estimates of adjustment for a sick or disabled child. In using such scales, many studies have not specifically assessed peer relationships.

Few studies have used multiple informants or multiple methods and few studies have used child, or peer reports. The main method of assessment has relied upon parental ratings, usually from mothers, and sometimes teacher ratings. Direct assessment of peer relationships is rare with multiscale diagnostic tools such as the CBCL dominating the method of assessment (Thompson and Gustafson, 1996).

A lack of focus on developmental stages has led to the use of samples of children that vary dramatically in terms of age. Relevant age and developmental differences may be obscured if studies do not specifically focus on these factors. Several reviewers have also pointed out that the frequent practice of comparing the adjustment of children with physical conditions to the adjustment of their healthy peers is problematic (Harper, 1991; Lemanek, 1994). These reviewers suggest that such comparisons promote a "difference/deficit" approach to the study of adjustment and do little to identify factors that promote positive adjustment within the population.

Despite such methodological concerns, the suggestion of peer relationship vulnerability within this population demands scrutiny. A most productive and positive line of enquiry would appear to be towards examining those parameters that are related to successful adjustment (La Greca, 1990). Sophisticated theoretical models have led to the study and identification of variables that may be seen to protect and assist the child on the path to successful adaptation and positive development (for examples see Wallander and Varni, 1992; Thompson and Gustafson et al, 1994). While the roles of social-cognitive variables such as self-perceptions have been extensively studied, the role of self-efficacy has not received the same attention. This is surprising considering the extensive application of the construct in other areas. In particular the study of self-efficacy has been prominent in the study of learning behaviours, health promoting behaviours (see Bandura, 1997) and has been identified in resilience literature as a factor that differentiates children who survive difficult life trajectories from those who succumb to risk (Rutter, 1985; Werner, 1993).

Existing social self-efficacy literature

Despite the wide application of self-efficacy theory both Bandura (1997) and Pajares (1997) have highlighted significant conceptual and measurement problems in the application of the construct. In order to adequately conceptualise self-efficacy measurement in the social domain it is important to first consider that self-efficacy beliefs are thought to vary across three dimensions, namely, level, strength and generality (Pajares, 1997). In considering the measurement of efficacy, the researcher must consider that there are different levels of task demand in any domain (Pajares, 1997). In the social domain, one may consider that talking to a group of friends is a relatively easy task to perform with very low level of demand. On the other hand striking up a conversation with an unfamiliar group may represent a higher level of skill demand. Self-efficacy items should attempt to assess the range of task demands in any given domain. Although some self-efficacy measures present tasks of increasing levels of demand (e.g. mathematics items) it is worth bearing in mind that some social tasks may not be so simply stratified. Bandura (1997) points out that some scales may quite appropriately include a range of items without specifying any particular order due to the idiosyncratic nature of task demands on any individual. In the social domain, the level of difficulty of any particular demand may be very subjectively determined.

Self-efficacy beliefs also vary in strength. The greater or stronger one's belief in their efficacy may lead to greater perseverance to perform a chosen activity and perform it successfully. While weak beliefs may be easily worn away, strong beliefs may be more robust to challenges (Bandura, 1997). For this reason, Bandura (1997) emphasises the need to assess efficacy on a scale that allows for wide differentiation of self-efficacy beliefs. The usual method of assessment uses a scale ranging from 0 to 100 indicating increasing level of certainty that one can do the task. Intervals of 0 to 10 have also been used but it is cautioned that using fewer steps risks some loss in differentiation (Bandura, 1997).

A major criticism of efficacy measurement has related to the issue of generality and correspondence with the criterial task. Generality typically refers to a level of assessment that is closely linked to domains of activity and situational contexts. In this way, research can assess the degree of generality of one's beliefs. Correspondence with the criterial task similarly refers to measurement which is consistent with the criterial task (Pajares, 1997). Global rather than specific forms of self-efficacy measurement include those that seek a judgement of one's general competence without specifying what exactly is being assessed. These type of measures imply that self-efficacy is a decontextualised judgement rather than a context-specific judgement. Pajares (1996) points out that generalised assessments of efficacy, such as a general academic efficacy scale for example, require people to make judgements of competence without any clear academic task in mind. In order to achieve an optimal level of self-efficacy assessment both Pajares (1996) and Bandura (1997) suggest that self-efficacy judgements should be consistent with the domain of functioning under investigation.

Ultimately, Bandura (1997) suggests that the adequacy of a self-efficacy measure can be assessed in terms of construct validity, specificity and range of task demands. The measure should demonstrate that it can measure what it claims to measure. To this end the measure should be able to demonstrate that it can predict the "...effects specified by the social cognitive theory in which the efficacy factor is embedded." (Bandura, 1997, p 45).

Despite the attention given to the role of self-efficacy in promoting positive academic behaviours it has been much less frequently examined in social contexts. Few studies have been located which examine, test or use theories or measures of social self-efficacy. Wheeler and Ladd (1982) pointed out that conceptualising social competence as skillful social behaviour closely aligns with Bandura's (1977) theory of self-efficacy- the belief that one can successfully perform behaviours required to produce desired outcomes.

Adopting this social-cognitive theory a small number of researchers including Wheeler and Ladd (1982) have examined the role of social forms of efficacy in promoting peer interactions (Wheeler & Ladd, 1982), in promoting social adjustment (Connolly, 1989) and more recently the relationship between social self-efficacy and academic self-efficacy (Patrick, Hicks and Ryan, 1997).

The Children's Self-Efficacy for Peer Interaction Scale (CSPI) developed by Wheeler and Ladd (1982) was designed to test 3rd to 5th grade children's perceptions of their ability to carry out prosocial, persuasive actions in specific peer situations. The items describe specific social situations such as joining a game and acting assertively in hostile peer situations. The authors confirmed that the measure had adequate psychometric properties and some construct validity in that it related to other measures of peer status and interactions. However, it is questionable whether the measure adequately distinguishes the level of belief in an ability to perform an action. The measure asks children to rate on a four point scale how hard or easy it would be to do a particular activity. This level of questioning is very restricted in the extent to which it may assess the strength in the belief of ability to execute a specific action. Similarly, although each item is specific about the social task, it does not explicitly ask about the strength of belief in ability to perform the task. Rather the scale implicitly refers to strength of belief.

The Connolly (1989) measure, the Social Self-Efficacy Scale does appear to more adequately assess levels of efficacy beliefs. The 7 point scale ranges from "impossible to do" to "extremely easy to do". Again, the judgement of skill is implicit rather than explicit in that the description of the task demand is not followed by a question referring to the rating of skill to perform the task. The level of difficulties of the tasks is also not specified by the researcher. Indeed several of the items appear to be tapping into very similar types of social tasks with similar level of task demand. For example, items concerning the ability to "ask someone over to your house" or "ask someone to go to a movie with you" appear to have similar levels of difficulty. Some of the items do not clearly specify any actual task, such as "Be involved in group activities", where the meaning of "involvement" is not clear and could encompass several different social skills.

The most recent of social self-efficacy scales located for this report appears to be that of Patrick et al (1997). The scale requires students to rate how true it is that they could perform a social task, such as starting a conversation with a classmate. While the task itself is quite specific the level of measurement does not appear to clearly tap the strength of one's belief of ability in the manner described by Bandura (1997). The use of a five point scale ranging from "not all true" to "very true" would not appear to adequately differentiate levels of belief to perform the given activity.

The role of self-efficacy in promoting adaptive behaviours.

The need to consider the role of social forms of self-efficacy among students with chronic conditions is particularly clear if one considers the sources and effects of self-efficacy beliefs. Sources and effects of efficacy may be considered in terms of four sources: enactive mastery experiences, vicarious experience, verbal and social persuasion and physiological states (Bandura 1997). In this framework, the emphasis is placed on the individual's cognitive interpretation of these factors that determines how they will be represented and reflected upon.

The life experiences of children with physical disabilities and other chronic medical conditions may create unique experiences that are not generally considered in the development of efficacy in other children. In particular the impact of illness parameters and constant onslaughts to the physiological state of the individual are not least of which need to be considered. However in order to speak to specific factors which relate to the social adjustment of the individual we must consider how the daily life experiences of these children differ from most children we may teach in our classrooms.

Many children with chronic conditions require frequent and prolonged periods of hospitalisation. School life is disrupted, with certain illness groups experiencing significant periods of absence. The nature of illness itself may induce states of physical disablement ranging from fatigue to complete immobility. Fatigue, pain and immobility may limit the child's usual peer interactions both in the classroom and playground. It is not surprising that several authors have suggested that this restriction to physical activity may pose a serious threat to children's peer relations (La Greca, 1990). Importantly, it is worth considering that the changing nature of children's peer interactions with age may mean that the ability to play physically may be more salient in some developmental periods than others.

Each of these experiences may affect the enactive mastery experience of the child. In some cases, individuals may miss out on the opportunity to engage in normative peer-based activities and thus the chance to practice and master appropriate social or academic skills. It is even likely that these children are required to practice and master quite different social skills than their peers. For example, La Greca (1990) suggests that telling other children about one's condition or explaining periods of illness may require skills not usually needed in everyday peer relationships.

Quite apart from the nature of the condition, the type of environment in which children gain experience must shape the enactive mastery experience of the student. A number of school practices shape the environment of children with disabilities in ways that may not be experienced by their classmates. For example, the nature and level of school integration determines the peer group in which the individual gains experience. A child integrated into a mainstream classroom will experience a peer group similar to any other school student. A student who is integrated but placed in a support unit however, will experience a peer group containing other children with disabilities as well as non-disabled peers. These differing peer groups may afford the student different peer interaction opportunities and different social reference groups. Similarly the role of the school in making school-based activities available to the individual ultimately determines the breadth and level of experience the child can hope to attain in these areas.

Vicarious experience in the school and social environment may also affect the self-efficacy of children with physical conditions. It has been well documented (Bandura, 1997) that efficacy beliefs can be enhanced or diminished by normative comparisons with the performance of others. While little is known about the social comparison processes of children with disabilities it must be assumed that these children engage in social comparison to a similar degree as other children. As referred to previously, it is likely that in the mainstream school environment the likely frame of reference will be an able-bodied peer group. Alternatively, the reference group in a supported setting may be children with other similar conditions. These differing reference groups may provide different points of comparison. To illustrate this point, McMaugh and Bowes (1995) noted that disabled children in mainstream settings referred almost entirely to able-bodied children on a measure of social comparisons of self-perception. Interestingly in their self-perceptions, most of these children attributed areas of poor performance to their disability. Case studies of children in supported settings where the primary social reference group was peers with similar conditions did not display a tendency toward either these comparisons or attributions. These differing points of comparison may make some children focus more upon their disability as a point of difference and potentially as a deficit.

In a recent study (in progress) children have reported acute awareness of their self in comparison to their peers. In one case a student was denied the opportunity to compete academically for an award due to being hospitalised during the testing period. The student and parents repeatedly asked for all schoolwork and the tests to be sent to the child. The tests and most schoolwork were never sent to the student. The school decided however, to give the student an award despite her not completing the series of tasks. The student's peers pointed out the student did not deserve to be rewarded for something she did not do. Consequently, the academic self-perception of the individual was severely diminished by the removal of an opportunity to compete and succeed on equal terms with her peers.

Bandura (1997) suggests that social evaluations of ability are conveyed indirectly and subtly to people believed to be of limited ability. Persuasive social feedback given either directly in the form of evaluation or indirectly in the form of social customs can undermine one's sense of efficacy or boost it. Fortunately, the supportive school environment provided to some students with disabilities probably acts as a positive source of social persuasion. On the other hand, negative peer interactions and school experiences such as the one mentioned above can undermine efficacy.

Despite the possibility of different peer opportunities, children with chronic conditions must still return to the regular school environment and succeed in the peer milieu in order to develop a positive social self-efficacy. More importantly, it is probable that the majority of children with chronic conditions are doing exactly this. As pointed out by La Greca (1990) it is more remarkable that so many children are positively adjusted than is the questionable possibility that some are not. For these reasons La Greca (1990), Lemanek (1994) and Harper (1991) among others, have called for a greater research effort to be expended upon identifying those parameters associated with positive adjustment within this population. The area of social adjustment research will benefit significantly from such an approach.

This study attempts to provide an additional insight into the role of social cognitive processes by examining the role of a form of social self-efficacy. This study adopts the measurement procedures outlined by Bandura (1997) and as such clearly assesses self-efficacy as a unique judgement of one's ability to execute specific social tasks. Similarly, the validity of the measure is assessed through its correspondence to assessment of social status among one's peers.

Method

Participants

An initial pilot study was conducted with a sample of 63 Year 6 students from a Catholic Primary school in Sydney. Following this, a second study using a slightly modified version of the instrument has commenced with Year 6 students from NSW Department of School Education sites located throughout metropolitan Sydney (n=170). In a third study, the instrument has been administered to several students with chronic physical conditions (n= 14). Children with chronic physical conditions were recruited through DSE lists of students receiving integration funding and specialist help. In addition, several students were contacted with the assistance of various support organisations for people with disabilities. The testing of the instrument with this group of students is continuing as a part of an ongoing research study. For all samples, there were approximately equal numbers of males and females and a mean age of 12 years. Written parental consent was obtained for all students.

Measures

During the pilot study only the Self-efficacy for Peer Interactions scale (SEPI) was administered. In subsequent studies, both the SEPI and a peer-based assessment were administered to all class groups and the individual students with chronic conditions. The peer assessment measure in these samples acted as an initial and partial outcome measure of validity for the SEPI.

The SEPI. The SEPI was developed to assess children's beliefs in their ability to carry out basic social tasks in order to achieve a social goal. An item pool was developed by surveying existing self-report social self-efficacy measures (Wheeler & Ladd, 1982; Connolly, 1989) and by reviewing the extensive body of literature on children's friendships and peer relationships. In particular literature on children's self-reported strategies for making friends (Wentzel & Erdley, 1993) guided the construction of items. In general, literature has identified that prosocial, co-operative behaviour tends to predict peer acceptance and positive peer status. On the other hand, a lack of prosocial skills and the presence of aggressive behaviour can lead to rejection and low acceptance. The item content in this measure focussed largely upon pro-social behaviours. The 10 items included in the scale were written to depict various scenarios that might be seen to occur in the school environment. The scenarios were written specifically with the pre to early adolescent group in mind to capture this group's tendency towards group and clique peer-relationships and a growing interest in forming closer, intimate relationships. The items were worded to refer broadly to the student's peer group rather than to a friendship group as it is reasonable to suspect than most children could comfortably carry out most social tasks with other children they feel closest to.

The SEPI was developed in an attempt to closely conform to the guidelines suggested by Pajares (1996; 1997) and Bandura (1997) regarding the specificity of self-efficacy assessment, and correspondence with criterial tasks. Both Bandura and Pajares have noted that self-efficacy research suffers from a problem of mismeasurement. This mismeasurement has been described by Pajares (1996) as resulting in self-efficacy assessments that have been too broad and general, lacking specificity of measurement and consistency with criterial tasks. These problems have led to ambiguity and misunderstanding about the potential role of self-efficacy beliefs in predicting performance outcomes. If for example one wished to examine the role of specific behaviours with certain specific outcomes then the nature of the self-efficacy instrument would need to reflect this level of specificity. Pajares (1996) points out that the optimal level of specificity must be determined by the complexity of the performance criteria with which it is being compared.

In developing the SEPI an attempt was made to develop items that reflected the specific context of peer interactions in the school environment. While being mindful of specificity, care was taken nevertheless, not to make each item so microscopically specific so as to lose practical utility. For example, items depict specific scenarios in the broader peer context. The aim of the measure is to assess peer interactions in the peer milieu generally, not within a certain group of children specifically. Similarly, basic social skills are depicted, such as being able to join in a conversation or group, rather than joining in a conversation about something particular or joining a specific group. The behaviour under scrutiny is the child's perception of his or her ability to accomplish basic social tasks that are important in the development of positive peer relationships.

In the manner recommended by Bandura (1997), the student is presented with the various task demands described in the form of a short scenario. The scenario is followed by the actual question item, which seeks a rating of "to what extent" the students believe that they "have the skill or ability" to perform the task. All scenarios are worded in the immediate present tense to emphasise the assessment of the student's present belief in their ability. The ten items are rated by the student on a scale ranging from 0 to 10 in equally spaced intervals, and anchored at the lower end by the words "Not at all sure" and at the upper end by "Totally sure." This judgement of present ability is conceptually different from an item that would ask about the possibility that the student will or might do the task in the indefinite future. This is an important point because perceived self-efficacy is not the intention to do a task. Rather it is the perception of capability that may lead to the intention to carry out the task. The responses to the items are summed to provide a total self-efficacy for peer interactions score with a possible range of 0-100.

An attempt was made to conceptually arrange items in order of an increasing level of difficulty. An item relating to "joining in a game you really like" with familiar peers preceded the item of "joining a different group for lunch" in the absence of familiar peers. It is questionable however whether or not individual students uniformly believe in the order of difficulty of tasks due to the idiosyncratic impact of particular task demands on any one individual (Bandura, 1997). The SEPI has however, included a range of task demands that conceptually vary in level of difficulty. Social demands such as negotiating peer group entry (especially with unfamiliar peers) have been identified by Dodge et al (1985) in the development of the Taxonomy of Problematic Social Situations for Children (TOPS) and La Greca (1992) in the development of the Social Anxiety Scale for Children (SACS) as potentially problematic and difficult social tasks for children to accomplish.

Peer Assessment

The peer assessment task used the rating scale procedures described by Asher and Dodge (1986). Children are given a list of all peers in their classroom. The students are then typically asked how much they "...like to play with..." each peer by placing a circle on a number line ranging from 1 to 5. A 1 response typically indicates "I don't like to.", and a 5 response indicates "I like to a lot." The measure used in this study adopted this procedure however the words "How much do you like to hang around with this person....", replaced "....play with..." in order to appeal to older and adolescent students, and the words "hardly ever" and "a lot" replaced "I don't like to" and "I like to a lot.", in order to avoid the connotation of stating one's dislike for a peer. The status classification scores are computed in the manner described by Asher and Dodge (1986) where the frequencies of 1 and 5 nominations are computed and transformed into standardised disliking and liking scores within each class group. This classification scheme yields 5 status groups including rejected, neglected, controversial, average and popular groups.

Procedure

The two measures were simultaneously group administered. Students were informed that they were participating in a study of children's friendships and peer interactions. Specific instructions for each scale were read aloud to the group with each student reading their own copy of the scale instructions. The group reviewed a sample item with fictional responses. Students were then allowed to practice with this sample item. Students completed the scale individually from this point onward, progressing at an individual rate. Specific instructions were given to any student with reading or language difficulties. The researcher rather than a teacher provided specific help when needed.

The pilot data (n=63) and research data (n=170) samples were collected in Term 4 of the school year with further administrations planned in the next year with a cohort of Year 7 students and a specific follow-up study with the individuals with chronic conditions.

Results

SEPI Internal consistency

A reliability analysis was conducted on the initial pilot study responses (n=63) to the SEPI scale to determine the internal consistency of responses. The Cronbach's alpha obtained for this initial sample was 0.86. This analysis was repeated on the full sample of 170 responses and revealed a Cronbach's alpha of 0.72. Both estimates are similar to the Cronbach's alpha reported in previous social self-efficacy studies (for example, Wheeler & Ladd, 1982, 0.82; Connolly, 1989, 0.9 to 0.95; Patrick, 1997, 0.72).

Relationship of SEPI to Peer status

To examine the relationship between the SEPI scores and peer status, a Pearson product-moment correlation analysis was performed on the data for a sample of the original 170 students for whom complete peer status data were available (n=82). This indicated a significant positive correlation between self-efficacy and peer status scores (r(82) = 0.25, p = 0.02), indicating that higher self-efficacy scores were associated with higher peer status.

Gender and condition differences

To determine whether total self-efficacy scores differed across boys and girls and across condition status, a 2 (gender) by 2 (condition status: chronic condition or no chronic condition) analysis of variance (ANOVA) was performed with data from the total sample (n=170).

An initial examination of conformity to ANOVA assumptions produced satisfactory results. The test for heterogeneity of variance was non-significant (Bartlett-Box F(3,2601) = 2.57, p > 0.05), as were the Kolmogorov-Smirnov (K-S) tests for non-normality in three of the cells (ps ( 0.07). In the fourth cell (females in the non-chronic illness group), the K-S statistic was significant (p < 0.01). However, the frequency distribution of scores in this cell indicated that this violation was due to skewness rather than to kurtosis. As Monte Carlo evaluations have indicated that the F test is robust to violations of skewness (Tabachnick & Fidell, 1996), although not of kurtosis, the effect of this departure on the analysis outcomes would be minimal.

Mean and standard deviations for total scores in each of the four cells of the design are shown in Table 1. Descriptive results for individual items are also shown to indicate variation across the four status groups for individual items. Relative difficulty rankings computed on the full 173-member sample (i.e., rather than for cell groups) for the 10 scale items indicated that Question 6 was found to be the least difficult (M = 9.1, SD = 7.9) followed by Question 8 (M = 8.4, SD = 2.0), Question 4 (M = 8.2, SD = 2.1), Question 3 (M = 8.1, SD =2.1), Question 7 (M = 7.9, SD = 2.2), Question 1 (M = 7.3, SD = 2.3), Question 10 (M = 7.2, SD = 2.6), and Question 5 (M = 7.1, SD = 2.7). Question 2 (M = 6.4, SD = 2.7) was rated as most difficult.

Table 1. Means and Standard Deviations for Total and Individual Item Scores by Gender and Condition Group Measure Females-Non-Chronic Males-Non-Chronic Females-Chronic Males-Chronic Total Scores 75.7(19.6) 78.0(16.2) 89.0(2.6) 62.0(16.9) Question 1 6.9(2.4) 7.8(2.0) 9.0(1.7) 6.6(2.4) Question 2 6.3(2.6) 6.6(2.6) 9.0(1.0) 5.4(4.0) Question 3 7.8(2.2) 8.5(1.8) 9.3(0.6) 7.8(2.9) Question 4 8.1(2.2) 8.4(2.0) 10.0(0.0) 6.6(2.7) Question 5 7.1(2.7) 7.3(2.5) 5.7(4.9) 5.5(2.8) Question 6 9.6(11.4) 8.6(2.1) 10.0(0.0) 8.5(1.8) Question 7 8.0(1.9) 7.9(2.4) 8.0(1.7) 6.0(2.9) Question 8 8.8(1.4) 8.2(2.2) 10.0(0.0) 6.1(2.5) Question 9 6.2(2.7) 7.2(2.5) 8.7(1.5) 4.4(3.6) Question 10 7.2(2.4) 7.4(2.7) 9.3(1.2) 5.1(2.9)

The ANOVA on these scores indicated no significant main effects for condition status (F(1,169) = 2.32, p > 0.05) or gender (F(1,69) < 1), and no significant gender by condition interaction effect (F(1,169) = 1.61, p > 0.05). Thus, there were no overall differences in mean SEPI scores across males and females or across condition groups. There were also no significant differences between males and females within either of the condition groups. It should be noted, however, that this outcome may reflect the small sample size involved in the present study, and the resulting low statistical power of the tests performed.

Discussion

The results presented here represent the preliminary stages of an ongoing investigation of the social adjustment processes of students with chronic physical conditions. The use of students without chronic conditions served the purpose of obtaining a large enough sample size to examine the psychometric properties of the instrument. As such these results appear to indicate that the Self-Efficacy for Peer Interactions Scale has reasonable internal consistency and appears to distinguish between a range of self-efficacy scores.

Bandura (1997) has questioned the purpose of trying to obtain high internal consistency among self-efficacy items as this may lead to a sacrifice of scale validity. Items that correlate too highly with one another may suggest that the self-efficacy measure is redundantly measuring only a small segment of the domain it claims to be assessing. If an efficacy scale displays an adequate range of task demands then differing responses to different levels of demands could be expected. Similarly, considering reliability as a test of stability over time may also be improper, as efficacy beliefs do not necessarily remain stable over time (Bandura, 1997, pp44- 45). Rather than relying on estimates of reliability, it is suggested that researchers rely on a sound conceptual scheme in the development of self-efficacy items. As such, the SEPI has not been subject to estimates of test-retest reliability, rather the notion that self-efficacy beliefs do not remain stable over time is being examined. Future work with the SEPI will examine whether or not efficacy beliefs change over time in contexts where the peer group has changed significantly, such as after the transition to high school.

The investigation of the perceived level of difficulty of each task revealed that student's perceptions of difficulty, as revealed by the average score for each item, differed to the conceptual assumption of level of difficulty. Item six concerning the ability to "offer your help" to other students received the highest ratings of ability to perform the task, while item two relating to joining in a conversation with classmates received the lowest ratings of ability to perform the task. These results seem to indicate that the tasks do vary in level of difficulty, however the basis on which these tasks vary is less clear. It seems likely that the students' conceptual reasons for rating the items in this way will prove the only explanation for such ordering. Future research will continue to examine the ordering of items.

The main test of the validity of a self-efficacy measure lies in its ability to measure what it purports to measure, an appropriate level of specificity and appropriate task demands. The measure gains validity as it successfully demonstrates the effects suggested in the social cognitive theory from which it developed (Bandura, 1997). The main test of validity that the SEPI was subject to in this investigation was in its relationship with peer social status. This investigation examined whether perceptions of efficacy for peer interactions was related to levels of peer status. The assumption here is that, conceptually at least, the ability to perform positive peer interactions should produce positive peer status outcomes.

This investigation yielded a moderate level of correlation between the SEPI and peer status, which was statistically significant. In this investigation, self-efficacy scores tended to increase as peer status scores increased. It may be suggested that this magnitude of correlation is not surprising due to the sheer complexity of peer relationships. A self-efficacy measure or indeed any single measure may, not easily explain the reasons why some children are accepted and others are not. Indeed such an explanation is not the aim of this research project.

The complex and ambiguous nature of social relationships indicates that assessing self-efficacy in the social domain may never be a straightforward task. Where the relevance of a particular form of mathematics self-efficacy to a particular outcome is relatively transparent, the relevance of any particular social behaviour to a particular outcome is not so clear. The reasons why their peers do not accept some children cannot simply be explained by their failure to master appropriate social behaviours. Therefore, the generality of specific social efficacy to specific social outcomes may not always be apparent.

In the context of the current research, the finding that self-efficacy scores do not appear to differ according to condition status is particularly interesting. The results are not indicative of differential reports of efficacy between condition and non-condition groups. Indeed, the lowest rating of self-efficacy on this measure to date was revealed by a student in the regular Year 6 class cohort. Not surprisingly this student was also rejected by his peers.

This early suggestion of no overall differences between the condition and non-condition status groups provides early support for the contention of Harper (1991) and others that preoccupation with a "difference/deficit" approach may be inappropriate. Future research with the complete group of students with chronic conditions will reveal whether this similarity remains even in the face of challenges to self-efficacy such as the transition to high school.

It is reasonable to speculate that transitional periods, although placing stress on most individuals, may place additional stress on students with physical conditions. If students with physical conditions do find it difficult to establish peer relationships, then self-efficacy may be challenged at this time. A more comprehensive data set will also allow for within-group variability in self-efficacy scores to be compared with other dimensions of self-beliefs such as social self-concept and perceptions of social support from one's peers and close friends. This data will also provide evidence of the generality of self-efficacy for peer interactions across other domains of social self-beliefs. Further tests of the validity of the SEPI will also include its correspondence with other indices of social adjustment including feelings of loneliness and social satisfaction.

This initial investigation provides a positive indication of the potential of the SEPI to provide an assessment of self-efficacy in social contexts. Importantly the SEPI can be considered an appropriate tool to use in a population of children with chronic physical conditions, as the tasks appear to be independent of physical abilities. The ability to successfully interact with one's peers and develop positive peer relations are critical elements of any child's development. This form of measurement allows for the examination of an additional facet of this complex domain of social functioning. Examining the potential of the self-efficacy construct to illuminate the social world of children should provide a fruitful and fascinating line of investigation for researchers in the future.

Acknowledgements: The authors would like to acknowledge and thank Dr Elaine Chapman and Ms Helen Watt for their statistical advice, support and helpful comments.

References

Achenbach, T. & Edelbrock, C. (1983) Manual for the Child Behavior Checklist and Revised Child Behavior Profile. Queen City Printers: Burlington, VT.

Asher,S. & Dodge, K. (1986) Identifying children who are rejected by their peers. Developmental Psychology, 22, 444-449

Bandura, A. (1977) "Self-Efficacy: Toward a unified theory of behavioural change." Psychological Review, 84, 191-215.

Bandura, A. (1997) Self-Efficacy: The Exercise of Control. New York, W.H. Freeman and Co.

Connolly, J. (1989) "Social self-efficacy in adolescence: Relations with self-concept, social adjustment and mental health." Canadian Journal of Behavioural Science, 21(3), 258-269

Dodge, K., McClaskey, C. & Feldman, E. (1985) Situational Approach to the Assessment of social competence in children. Journal of Consulting and Clinical Psychology, 53, 344-353.

Eiser, C. (1990) Psychological effects of chronic disease. Journal of Child Psychology and Psychiatry,31(1), 85-98.

Harper, D. (1991) Paradigms for investigating rehabilitation and adaption to childhood disability and chronic disease. Journal of Pediatric Psychology, 16, 533-542.

LaGreca, A., Kraslow-Dandes, S., Wick, P., Shaw, K., & Stone, W. (1992) Development of the social anxiety scale for children: Reliability and concurrent validity. Journal of Clinical Child Psychology, 17, 84-91

LaGreca, A. (1990) Social consequences of pediatric conditions: Fertile area for future investigation and intervention. Journal of Pediatric Psychology, 15, 285-307.

Lavigne, J., & Faier-Routman, J.(1992) Psychological adjustment to pediatric physical disorders: A meta-analytic review. Journal of Pediatrics, 17, 133-158.

Lemanek, K. (1994) Editorial: Research on pediatric chronic illness: New directions and recurrent confounds. Journal of Pediatric Psychology, 19, 143-148.

McMaugh, A. & Bowes, J. (1995) An investigation of the self-concept and social comparison processes of young adolescents with physical disabilities. A paper presented at the annual meeting of the Australian Association for Research in Education, Hobart, November, 1995

Nassau, J. & Drotar, D. (1995) "Social competence in children with IDDM and asthma: Child, teacher and parent reports of children's social adjustment, social performance and social skills." Journal of Pediatric Psychology, 20, 187-204.

Pajares, F. (1996) Current Directions in Self Research: Self-Efficacy. Paper presented at the annual meeting of the American Educational Research Association, New York, April, 1996.

Pajares, F. (1997) Current directions in self-efficacy research. In M. Maehr & P. Pintrich (Eds.) Advances in Motivation and Achievement, 10.

Patrick, H., Hicks, L., & Ryan, A. (1997) Relations of perceived social efficacy and social goal pursuit to self-efficacy for academic work. Journal of Early Adolescence, 17, 109-128.

Perrin, E., Stein, R., & Drotar, D. (1991) "Cautions in using the child behavior checklist: Observations based on research about children with a chronic illness." Journal of Pediatric Psychology, 16, 411-422.

Perrin, J. (1985) Introduction. In N. Hobbs & J. Perrin (Eds.) Issues in the care of children with chronic illness: A source book on problems, services and policies. (pp1-10) San Francisco: Jossey-Bass.

Rutter, M. (1985) "Resilience in the face of adversity: Protective factors and resistance to psychiatric disorder." British Journal of Psychiatry, 147, 598-611.

Spirito, A., DeLawyer, D., & Stark, L. (1991) "Peer Relations and social adjustment of chronically ill children and adolescents." Clinical Psychology Review, 11, 539-564.

Tabachnick, B. & Fidell, L. (1996) Using Multivariate Statistics. New York: HarperCollins College Publishers.

Thompson, R. & Gustafson, K. (1996) Adaptation to Chronic Childhood Illness. American Psychological Association: Washington.

Wallander, J. & Varni, J. (1992) Adjustment in children with chronic physical disorders: Programmatic research on a disability-stress-coping model. In A. LaGreca, L. Siegal, J. Wallander & C. Walker (Eds.) Stress and Coping with Pediatric Conditions, (pp279-298) New York: Guildford Press.

Wentzel, K. & Erdley, C. (1993) Strategies for making friends: Relations to social behaviour and peer acceptance in early adolescence. Developmental Psychology, 29, 819-826.

Werner,E. (1993) " Risk, Resilience and Recovery: Perspectives from the Kauai Longitudinal Study." Development and Psychopathology, 5, 503-515.

Wheeler, V. & Ladd, G. (1982) "Assessment of Children's Self-Efficacy for Social Interactions with Peers." Developmental Psychology, 18 (6), 795-805.