Gifted Child Quarterly - CiteSeerX

Gifted Child Quarterly http://gcq.sagepub.com

Social Self-Concept, Social Attributions, and Peer Relationships in Fourth, Fifth, and Sixth Graders Who Are Gifted Compared to High Achievers Sherry K. Bain and Sherry Mee Bell Gifted Child Quarterly 2004; 48; 167 DOI: 10.1177/001698620404800302 The online version of this article can be found at: http://gcq.sagepub.com/cgi/content/abstract/48/3/167

Published by: http://www.sagepublications.com

On behalf of:

National Association for Gifted Children

Additional services and information for Gifted Child Quarterly can be found at: Email Alerts: http://gcq.sagepub.com/cgi/alerts Subscriptions: http://gcq.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 17 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://gcq.sagepub.com/cgi/content/refs/48/3/167

Downloaded from http://gcq.sagepub.com at PENNSYLVANIA STATE UNIV on February 5, 2008 © 2004 National Association for Gifted Children. All rights reserved. Not for commercial use or unauthorized distribution.

S O C I A L / E M OT I O N A L S TAT U S O F G I F T E D C H I L D R E N

Social Self-Concept, Social Attributions, and Peer Relationships in Fourth, Fifth, and Sixth Graders Who Are Gifted Compared to High Achievers Sherry K. Bain Sherry Mee Bell University of Tennessee

ABSTRACT Socially related self-concept, attributions for social success and failure, and peer relationships were investigated for fourth through sixth graders identified as intellectually gifted and a comparison group of high achievers not identified as gifted. The group identified as gifted scored significantly higher on 3 of 4 socially related self-concept subscales (p < .01) and on the general self-concept scale (p < .05) of the Self-Description Questionnaire–I (Marsh, 1988). On the Student Social Attribution Scale (Bell & McCallum, 1995), an interaction effect was noted, with the group identified as gifted scoring higher than the comparison group on attributions for social success due to ability and effort and lower on attributions for social failure due to ability, effort, and task difficulty. Though teachers‘ ratings of peer relationships yielded no differences between the 2 groups, boys across groups were rated significantly higher than girls (p < .05). The purpose of this study was to investigate social self-concept, attributions for social success and failure, and peer relationships of children identified as intellectually gifted compared to high achievers not identified as gifted. A frequently voiced perception of the general population is that individuals who are gifted suffer from unusual emotional and social vulnerabilities. This perception can be linked to historically prominent concerns voiced by authors such as Lombroso (1895) and by more recent authors (Lajoie & Shore, 1981). Anecdotal reports also raise concerns, such as the 1981 report of the suicide of a gifted high school student (as reported in Neihart, 1999). College students have reported perceptions that individuals who are gifted have a higher probability of problems in the social realm (Solano, 1987). Rimm

(2002) pointed out that gifted adolescents often report feeling “different” from peers and argued that this feeling of being different may affect their social relationships. Anecdotal reports, popular perceptions, and limited evidence from the professional literature help to sustain the impression that children who are gifted indeed constitute a special group in terms of social and emotional needs. However, there is some evidence that, as a group, people who are gifted tend to exhibit stronger emotional or personal adjustment than the general population.

PUTTING THE RESEARCH TO USE Results from this study add to the body of literature providing evidence that children identified as gifted, as a group, are no more vulnerable to socially related self-concept problems than the general population or, in this case, a population of high-achieving students. The group identified as gifted scored higher on several self-concept scales than the high-achieving comparison group and had a stronger tendency to attribute social successes to ability and effort, rather than luck or task difficulty. Implications of the results for educational purposes are to concentrate on the social strengths of children in gifted programs who are functioning well in their school environments and to encourage social interactions with children across the regular education spectrum to develop these skills. However, for the student who is gifted, but truly has problems in the area of social/emotional functioning, teachers and parents are advised to seek evaluation and services through special education programs or through Section 504 provisions for students with disabilities.

G I F T E D C H I L D Q U A RT E R LY • S U M M E R 2 0 0 4 • VO L 4 8 Downloaded from http://gcq.sagepub.com at PENNSYLVANIA STATE UNIV on February 5, 2008 © 2004 National Association for Gifted Children. All rights reserved. Not for commercial use or unauthorized distribution.

NO 3

167


Terman and Oden (1947) presented early evidence for lower incidences of mental illness and adjustment problems among gifted students than across the general population. Beer (1991), Grossberg and Cornell (1988), and Nail and Evans (1997) have provided additional evidence for higher levels of emotional adjustment among students identified as gifted. For recent reviews of the literature on the social and emotional development in children who are gifted, see Neihart (1999) and Neihart, Reis, Robinson, and Moon (2002). Despite evidence that students who are gifted are well adjusted, studies focusing specifically on the social domain (including self-concept, attributions, and peer relationships) for this group are fairly recent and few in number. In contrast, studies addressing general selfconcept, academic self-concept, and attributions for academic success and failure are plentiful (see Hoge & Renzulli, 1993; Robinson, 2002). Research dealing specifically with social self-concept, attributions for social success and failure, and peer relationships is described below.

Self-Concept Among studies examining self-concept in students who are gifted, three have included elements of socially related self-concept. Brody and Benbow (1986) administered measures of general self-esteem and social functioning to 173 adolescents who were highly talented in math or verbal abilities based on their performance on the Scholastic Aptitude Test (SAT) and to a comparison group of 205 adolescents with high academic performance. Students from the high academic performance group had scored at or above the 97th percentile on an in-grade test and were initial participants in talent searches with the students in the highly talented groups. Membership in the highly talented and high academic performance groups did not overlap. No differences were found between the two groups on the self-esteem measure. In contrast, Pyryt and Mendaglio (1994) reported a significant difference on the social domain of the PyrytMendaglio Self-Perception Survey for a Canadian sample of 45 eighth and ninth graders identified as gifted and 52 peers not identified as gifted. In this case, the children identified as gifted achieved a higher mean score than their peers. Finally, Ablard (1997) obtained mixed results in her comparison of 174 academically talented eighth graders 168

to normative groups using the Multidimensional SelfConcept Scale (MSCS; Bracken, 1992) and the Adjective Check List (ACL; Gough & Heilbrun, 1983). While no differences in social self-concept were noted on the MSCS, moderately higher general self-confidence scores were obtained by the academically talented group on the ACL. Gender differences between high-achieving females and the comparison group occurred on the subscales and are discussed later. Some researchers have theorized that educational placement inf luences the self-concepts of gifted children, akin to the Big-Fish-Little-Pond effect described by Marsh (1987). For instance, Colangelo and Davis (1991) suggested that external recognition, such as being identified as gifted, improves self-concept. In contrast, Coleman and Fults (1985) suggested that a social comparison process might cause a decline in self-concept when talented children are placed in a class of similarly talented children, causing a devaluation of abilities in comparison to other high-achieving peers. Conversely, when these students return to placement in a heterogeneous classroom, their self-concepts might be expected to increase. Lending empirical credence to the purported relationship between placement and self-concept, Kulik and Kulik’s (1992) meta-analysis indicated slightly negative self-perceptions for students identified as gifted and placed in homogeneous accelerated programs when compared to their nonparticipating intellectual peers. The authors suggested that the difference should be considered negligible, probably representing a drop to more realistic self-perceptions after a change in placement. Hoge and Renzulli (1993) also presented meta-analytic results of studies involving educational placement of children who are gifted, finding generally high self-concept for students who are gifted, with some variability across studies. Social self-concept was not addressed specifically in either meta-analysis. There is some evidence that females identified as gifted have lower self-concepts in specific areas than gifted males. For example, Ablard (1997) reported that academically talented females exhibited lower ACL scores in socially related subscales (e.g., Succorance, Abasement) than talented males. According to Reis (2002), females identified as gifted tend to experience diminished selfconfidence as they enter adolescence, likely due to parental, teacher, and societal expectations. Citing a 1991 study by Kline and Short, Reis noted that self-confidence of gifted girls decreased from elementary to secondary school.

G I F T E D C H I L D Q U A RT E R LY • S U M M E R 2 0 0 4 • VO L 4 8

NO 3



Attributions for Success and Failure Studies of attributions among children who are gifted are rare in the refereed literature. Brody and Benbow (1986) included general attributions in their comparison of highly talented young adolescents identified in talent searches based upon high SAT scores and high academic performers identified by in-grade percentile scores at the 97th percentile or above. Brody and Benbow found that the highly talented group exhibited a higher internal locus of control than the comparison group of high academic performers. Only four locus-of-control items represented the attribution scale without focusing on specific domains such as academics or social competence. More recently, Bell and Schindler (2002) examined the social and academic attributions of 270 fourth, fifth, and sixth graders using the Student Social Attribution Scale (Bell & McCallum, 1995) and the Student Academic Attribution Scale (Bell, McCallum, & Doucette, 2003). These researchers found no differences between children identified as gifted (n = 22) and students in regular education on attributions for social success and failure. In contrast, attributions for academic success and failure followed the trend reported elsewhere in the literature (see Robinson, 2002), with students identified as gifted reporting higher internal attributions for academic success than nongifted peers.

Peer Relationships In spite of popular assumptions regarding an inverse relationship between peer acceptance and giftedness (e.g., Solano, 1987), there is little evidence that children identified as gifted suffer from greater peer rejection than the general population. Characteristics common in popular children, including good social skills, few behavior problems, leadership skills, high academic success, and high self-esteem (Frentz, Gresham, & Elliott, 1991; Jackson & Bracken, 1998; Kennedy, 1990) are often noted in individuals who are gifted. These qualities can certainly have positive inf luences on peer acceptance, as well as on the development of self-concept, as demonstrated by Coie, Dodge, and Coppotelli (1982). According to Rimm (2002), no studies have supported the notion that gifted elementary-aged students are less popular than their peers. In fact, Rimm cited several studies suggesting gifted elementary-age students may be more popular than peers. Bell and Schindler

(2002) also provided support for this view, finding that children identified as gifted were rated significantly higher in popularity by peers than were their regular education classmates. Rimm noted, however, that this edge in popularity disappears in adolescence. She also discussed some gender differences in popularity, citing a 1990 study by Nichols in which gifted girls were found to be the least popular, behind gifted boys and nongifted boys and girls. Further, without supplying empirical evidence, Reis (2002) noted qualitatively that teachers tend to like girls described as smart less than other students. Contrary to evidence of generally strong peer relations for gifted children, Rimm (2002) has noted that these children often report feeling different from peers regardless of whether they receive gifted education services. Providing support for this notion, Brody and Benbow (1986) found that a group of highly talented students, based upon high SAT scores achieved during talent searches, rated themselves less popular than did a high-performance comparison group that had achieved in-grade test scores at or above the 97th percentile.

Summary of Previous Findings Although the consensus of public opinion supports the perception of gifted individuals being at exceptional risk for emotional or social difficulties, the empirical literature has frequently not supported this perception. Only three studies have addressed social self-concept among gifted students, all of which focused on the early adolescent years and yielded conf licting results. Metaanalyses of the effects of gifted placement on self-perceptions have also produced conf licting conclusions. Additionally, only two studies of the social attributions of giftedness are available, again yielding conf licting results. A single study of peer relationships has indicated that gifted children are rated higher in acceptance by peers, despite previous evidence that these children view themselves as less popular than their peers. Some gender differences in the area of peer relationships have been noted, particularly for adolescent age groups. Notably missing in the research literature are investigations examining the combination of self-concept, social attributions, and peer relationships of gifted students in preadolescence. Pyryt and Mendaglio (1994) suggested the need for additional research in the areas of “ref lected appraisals, social comparison, and attribution” (p. 303). The preadolescent age range, from approximately fourth through sixth grades, represents a period


NO 3

169


during which many public schools begin the process of identifying gifted children and placing them in special classrooms for at least part of the week. Aside from recognizing that they are receiving special treatment, some of these students may begin their first encounters with peers who interact at their intellectual level. In addition, late elementary and early middle school represents a time when it is suggested that gifted females may begin to experience a decrease in self-concept. Based upon these concerns, the present study was designed to focus on students from fourth to sixth grades.

Purpose of the Study The objective of this study was to investigate socially related self-concept, attributions for social success and failure, and peer relationships for fourth through sixth graders identified as gifted and a comparison group of students who are high achievers, but not identified as gifted. High-achieving students were chosen as the comparison group to eliminate gross achievement differences as a confounding variable. Hypothesized relationships were based upon a review of the empirically based literature discussed above. No significant differences were hypothesized between students identified as gifted and students identified as high achievers on the variables of social self-concept, social attributions, and peer relationships. Because differences noted in the literature between boys and girls who are gifted tend to be more robust in high school, no hypothesized differences based on gender were made for males and females in grades 4–6. However, if differences based on identification as gifted or on gender were found, especially if either group exhibited lower self-concepts than expected in the general population or unhealthy attribution patterns for social situations, considerations for screening and interventions for that group could be warranted.

Method

Procedure

Participants Participants were 93 elementary and middle school students: 29 fourth graders, 32 fifth graders, and 32 sixth graders. Forty-eight participants were male; 45 were female. Ninety participants were White, 2 were African American, and 1 was Asian American. Twenty-six participants (14 males, 12 females) across the three grade lev170

els were identified as intellectually gifted and received special education services under this category. Students in the intellectually gifted group had been identified based on state special education criteria, which include performance approximately 2 standard deviations above the mean on an individually administered intelligence test; academic achievement at or above the 96th percentile in reading, language, mathematics, science, or social studies on a group or individually administered test; and teacher ratings. The remaining 67 participants (33 males, 34 females), also spread across the three grade levels, were high achievers based upon composite achievement scores at the 85th percentile or above on the state-administered achievement test, the Terra Nova based upon the Comprehensive Test of Basic Skills (CTB/McGraw-Hill, 1996). The mean age of participants was 11.06 years; the range was 9.44 to 12.94 years (SD = .93). Permission was obtained from school officials and the appropriate institutional review board to carry out the study. Participants were recruited from one elementary school and one middle school from each of two districts within the same county in a southeastern state. The two districts have different policies for providing services for gifted students. District A, a larger, research-based community, identifies gifted children and places them in pull-out programs at the elementary level and in scheduled classes specifically for gifted children at the middleschool level. District B, a primarily rural community, does not identify gifted children and serves all children not identified with special needs in the regular classroom. Participants classified as high achievers were drawn from both District A (n = 38) and District B (n = 29), while participants classified as gifted (n = 26) were drawn from District A. Based upon the total population of children from District B in fourth, fifth, and sixth grade, we estimated that approximately 6 students from the group of high achievers might qualify for gifted identification and services if attending District A.

After permission was obtained from parents or guardians, two questionnaires, the Self-Description Questionnaire–I (SDQ-I; Marsh, 1988) and the Student Social Attribution Scale (SSAS; Bell & McCallum, 1995), were administered to participants in groups ranging in size from 6 to 10. Questionnaires were administered by the authors and advanced graduate students during the school day in a classroom set aside for the study. After


NO 3



participants gave oral assent, the questionnaires were then administered in counterbalanced order of presentation across groups. The Teacher’s Social Rating was distributed to the students’ regular or homeroom teacher on the same day, asking him or her to provide a rating of each student’s peer relationships.

Instruments Self-Description Questionnaire–I (SDQ-I). The SDQ-I (Marsh, 1988) measures eight components of preadolescent self-concept based on the Marsh-Shavelson model of multidimensional, hierarchical self-concept (Marsh & Shavelson, 1985). Normed on 3,562 Australian second through sixth graders from geographically and socioeconomically diverse backgrounds, the SDQ-I appears to be theoretically and empirically sound. According to Marsh (1988), exploratory and confirmatory factor analyses of the SDQ-I provide support for eight factors. Internal consistency coefficients for SDQ-I subscales were all above .8, and total scale reliabilities were above .9. Testretest reliability coefficients over a 6-month period averaged .61 for the subscales and .65 for the three total scales. Although no children from the United States were included in the normative sample, clinical data provide support for using the scale in the United States (Marsh). The SDQ-I measures two broad areas, academic and nonacademic, each with several subscales. In addition, a General-Self subscale (GSelf) measures a global, noncontext-specific self-concept. The SDQ-I yields T scores and percentiles for its eight subscales; however, because of a ceiling effect based on T scores for participants in the present study, raw scores were used in the data analysis. The socially related subscales, Physical Ability (Physical Ability), Physical Appearance (PhAp), Peer Relations (PeerRel), Parent Relations (ParRel), and the General Self (GSelf) subscale, were examined in this study. Student Social Attribution Scale (SSAS). The SSAS is a 30-item expansion of the SSAS developed by Bell and McCallum (1995). The SSAS assesses causal attributions for social success and failure in school-related situations. Students are presented with social situations, half with positive (i.e., “success”) outcomes and half with negative (i.e., “failure”) outcomes. They are asked to rate how likely each of one of four causal statements (representing ability, effort, chance, and task difficulty) is in this particular situation. Each causal statement is rated on a Likert scale (1 = seldom, 2 = sometimes, and 3 = often). An example from the SSAS is the following: “I am invited to a classmate’s party. It is because (a) A lot of people were

invited, (b) I make friends easily, (c) I work hard at making friends, and (d) Recently, making friends is easy for everyone.” The 1995 version of the SSAS, consisting of 12 item stems, produced a six-factor structure: Success Ability (SA), Success Effort (SE), Success External, Failure Ability (FA), Failure Effort (FE), and Failure External. In the 1995 version, both Chance and Task Difficulty items were included on one external scale, which likely contributed to the low retest reliabilities for both external scales (.24). Test-retest reliabilities for the internal scales ranged from .66 to .77. To improve psychometric properties, an expanded 30-item stem form, used in the present study, was developed. Item responses were revised to include two external attribution choices: chance/luck and task difficulty. This resulted in four additional subscales labeled Success Chance (SC), Success Task Difficulty (ST), Failure Chance (FC) and Failure Task Difficulty (FT). For the 30-item stem form, Cronbach’s alphas ranged from .76 to .93. Test-retest reliability data, gathered on a sample of 19 sixth graders with a 2-week time interval, produced generally strong reliabilities ranging from .74 to .84 across four internal and external subscales (Bain & Reece, 2002). For the purposes of the present study, the eight subscales examined were SA, SE, SC, ST, FA, FE, FC, and FT. Teacher’s Social Rating (TSR). The TSR was designed to obtain a gross estimate of each student’s relative functioning level in peer relationships. The preferred method for obtaining this information is through peer ratings based upon methods used by Coie and Dodge (1988). However, Coie and Dodge’s method, and variations that use classmates to identify popular and unpopular peers, can appear aversive to school administrators, teachers, and parents who provide consent. Therefore, the TSR was developed to provide an estimate of peer functioning from participants’ regular education teacher or homeroom teacher. The TSR asks the child’s teacher to rate him or her on a scale from 1 to 5 (1 = has very frequent positive interactions with peers or classmates, 2 = has a fairly high level of positive interactions with peers or classmates, 3 = has an average level of positive interactions with peers or classmates, 4 = interacts positively with peers and classmates on an occasional basis, 5 = generally avoids interactions with peers and classmates). Reliability for this instrument was investigated by asking four teachers in two of the participating schools to rate the same children 2 weeks after initial ratings. A perfect correspondence between ratings was obtained for the 20 children rated twice. Construct validity for the TSR is addressed later in this report.


NO 3

171


Table 1 Means, Standard Deviations, and F for Main Effects for Self-Description Questionnaire-I (SDQ-I) Subscales

Instrument Subscale/Scale

Gifted Mean, SD (n = 26)

High Achievers Mean, SD (n = 67)

F (df) for Main Effect for Placement

SDQ-I Physical Ability SDQ-I Physical Appearance SDQ-I Peer Relations SDQ-I Parent Relations SDQ-I General Self

50.77, 9.58 50.65, 8.17 49.00, 8.40 46.73, 10.65 51.69, 7.61

41.27, 10.76 41.28, 12.64 41.31, 11.85 43.40, 10.66 43.81, 11.70

15.46 (1, 89)** 11.75 (1, 89)** 8.73 (1, 89)** 1.78 (1, 89)NS 5.35 (1, 89)*

Note. *significance level < .05; **significance level < .01; NS = not significant.

Data Analysis Data were analyzed through two multivariate analyses of variance and one univariate analysis of variance. Independent variables for each of the three analyses were placement (intellectually gifted [G] or high achiever [HA]) and gender. Dependent variables for the first multivariate analysis were SDQ-I subscale raw scores: PhAb, PhAp, PeerRel, ParRel, and GSelf. Dependent variables for the second analysis were SSAS subscale raw scores: SA, SE, SC, ST, FA, FE, FC, and FT. The TSR was the dependent variable for the univariate analysis. An a priori significance level was set at .05 for the three analyses. Grade levels were not entered in analyses as independent variables because the difference in school levels (elementary vs. middle school) for fifth graders across districts was considered a confounding element. Following examination of multivariate results, interaction effects for SSAS attributions across social success and failure situations were examined based upon the constructs of ability, effort, chance, and task difficulty. Repeated measures analyses were carried out for each of the four constructs. Finally, a post-hoc correlation matrix was generated to examine relationships between SDQ-I scores and SSAS scores and to determine construct validity for the TRS.

Results No differences were hypothesized between gifted and high-achiever groups or between gender groups on social self-concept represented by the SDQ-I subscales (Physical Ability, Physical Appearance, Peer 172

Figure 1. Comparison of Self-Description Questionnaire–I (SDQ-I) subscale means for Physical Ability (PhAb), Physical Appearance (PhAp), Peer Relationships (PR), Parent Relationships (ParR), and General Self (GSelf) between gifted and high-achieving groups Relations, Parent Relations, and General Self), on social attributions represented by the SSAS subscales (Success Ability, Success Effort, Success Chance, Success Task Diff iculty, Failure Ability, Failure Effort, Failure Chance, and Failure Task Difficulty), or on peer relations represented by TSR scores. Results of two multivariate analyses investigating SDQ-I and SSAS differences failed to support the null hypothesis in general for placement groups, but supported it for gender groups. Results of the TSR analysis supported the null hypothesis for placement groups, but not for gender.


NO 3



Table 2 Means, Standard Deviations, and F for Main Effects for Student Social Attribution Scale (SSAS) Subscales Instrument Subscale/Scale

Gifted Mean, SD (n = 26)

High Achievers Mean, SD (n = 67)

SSAS Success Ability SSAS Success Effort SSAS Success Chance SSAS Success Task Difficulty SSAS Failure Ability SSAS Failure Effort SSAS Failure Chance SSAS Failure Task Difficulty

35.15, 5.35 39.58, 5.53 27.08, 5.11 29.35, 5.81 19.88, 3.78 18.92, 2.58 27.12, 5.22 25.69, 5.33

30.14, 7.22 31.44, 11.98 29.86, 6.08 28.88, 4.61 28.77, 10.97 28.61, 11.95 28.86, 5.22 29.27, 5.09

F (df) for Main Effect for Placement 10.54 (1, 88)** 11.14 (1, 88)** 4.10 (1, 88)* 0.18 (1, 88)NS 15.71 (1, 88)** 16.18 (1, 88)** 2.01 (1, 88)NS 8.52 (1, 88)**

Note. *significance level < .05; **significance level < .01; NS = not significant.

In the first multivariate analysis, with placement and gender as independent variables and SDQ-I subscales as dependent variables, a main effect for placement was indicated (Wilks’ Lambda F [5, 85] = 3.44; p < .01). Gender and interaction effects were not significant. Of the five SDQ-I subscales, Parent Relations was the only one that did not produce a significant difference. All significant SDQ-I differences favored the G group (see Figure 1). See Table 1 for SDQ-I subscale means, standard deviations, Fs, and significance levels. In the second multivariate analysis, with placement and gender as independent variables and SSAS subscale scores as dependent variables, a main effect was noted for placement (Wilks’ Lambda F [5, 81] = 2.49; p < .05) while gender and interaction effects were not significant. Differences on the SSAS subscales of Success Ability, Success Effort, Success Chance, Failure Ability, Failure Effort, and Failure Task Difficulty reached significance, with no significant differences on Success Task Difficulty and Failure Chance. An inspection of the direction of significant differences indicates variations in direction between the two groups, with the gifted group higher on the Success Ability and Success Effort subscales and the high-achiever group higher on Success Chance, Failure Ability, Failure Effort, and Failure Task Difficulty (see Figure 2). Table 2 lists SSAS means, standard deviations, subscales Fs, and significance levels. A univariate analysis of variance, with TSR scores as the dependent variable and placement and gender as independent variables, produced a nonsignificant main effect for placement, a significant main effect for gender,

Figure 2. Comparison of Student Social Attributions Scale subscale means for Success Ability (SA), Success Effort (SE), Success Chance (SC), Success Task Difficulty (ST), Failure Ability (FA), Failure Effort (FE), Failure Chance (FC), and Failure Task Difficulty (FT) between gifted and high-achieving groups and a nonsignificant interaction effect for placement and gender. Mean TSR scores for the gifted and high-achiever groups were 1.96 (SD = 1.09) and 1.88 (SD = 1.04), respectively. A main effect for gender was noted (Wilks’ Lambda F [1, 88] = 4.00; p = .05); males received a mean TSR score of 2.08 (SD = 1.13); females, 1.70 (SD = .93). An examination of multivariate results from the SSAS indicated possible interaction effects for attribution constructs based upon ability, effort, chance, and task difficulty. Post hoc repeated measures analyses


NO 3

173


Table 3 Results of Four Repeated Measures Analyses Comparing Student Social Attribution Scale (SSAS) Success and Failure Attribution Scores for Ability, Effort, Chance, and Task Difficulty With Placement (Gifted vs. High Achievers) as Independent Variable

SSAS Constructs Ability (SA & FA) Effort (SE & FE) Chance (SC & FC) Task Difficulty (ST & FT)

Main Effect for Construct Wilks’ Lambda F (df)

Interaction Effect Wilks’ Lambda F (df)

22.02 (1,90)** 24.78 (1,90)** 0.74 (1,90)NS 4.11 (1,90)*

15.38 (1,90)** 14.26 (1,90)** 0.87 (1,90)NS 6.34 (1,90)*

Note. SA = Success Ability, SE = Success Effort, SC = Success Chance, ST = Success Task Difficulty, FA = Failure Ability, FE = Failure Effort, FC = Failure Chance, FT = Failure Task Difficulty. *significance level < .05; **significance level < .01; NS = not significant

Figure 3. Interaction effect between gifted and highachieving groups for social attributions for ability

Figure 4. Interaction effect between gifted and highachieving groups for social attributions for effort

were completed designating placement as the independent variable and two levels of each construct (success levels and failure levels) as the repeated measures. That is, SA and FA were investigated as two levels of Attributions for Ability, Effort, Task Difficulty, and Chance. Three of the four repeated measures analyses yielded significant main effects for success versus failure for Attributions for Ability, Effort, and Task Difficulty, and significant interaction effects between success/failure and placement. The analysis investigating Attributions for Chance yielded nonsignificant main and interaction results (see Figures 3–6 and Table 3). Specifically, the gifted and high-achiever groups differed significantly on Success Ability, Failure Ability, Success Effort, Failure Effort, and Failure Task

Difficulty subscales, but not on the Success Task Difficulty subscale. Paired sample t-tests established significant within-group differences for the gifted group between Success Ability and Failure Ability (p < .01), Success Effort and Failure Effort (p < .01), and Success Task Difficulty and Failure Task Difficulty (p < .05). Within-group differences were not found in similar comparisons for the high-achiever group. Results of a correlation matrix pairing SDQ-I subscale scores and SSAS scores is presented in Table 4. Only the SSAS subscale Success Task difficulty failed to correlate significantly with SDQ-I subscales. The range of correlations for the remaining comparisons of SSAS and SDQ-I subscales was .33 to .83 in absolute terms, with negative correlations in expected directions. SSAS

174


NO 3



Figure 5. Interaction effect between gifted and highachieving groups for social attributions for chance

Figure 6. Interaction effect between gifted and high-achieving groups for social attributions for task difficulty

Table 4 Intercorrelations Between Self-Description Questionnaire-I (SDQ-I) and Student Social Attribution Scales (SSAS) Subscales SDQ-I Subscales SSAS Subscales SA SE SC ST FA FE FC FT

PhAb

PhAp

PeerRel

ParRel

Gself

.541* .631* -.416* .018 -.640* -.626* -.382* -.327*

.711* .783* -.579* -.016 -.829* -.824* -.478* -.523*

.703* .685* -.448* .063 -.781* -.719* -.454* -.530*

.539* .670* -.372* .081 -.640* -.609* -.419* -.345*

.635* .723* -.482* -.041 -719* -712* -392* -465*

Note. PhAb = Physical Ability, PhAp = Physical Appearance, PeerRel = Peer Relations, ParRel = Parent Relations, Gself = General Self, SA = Success Ability, SE = Success Effort, SC = Success Chance, ST = Success Task Difficulty, FA = Failure Ability, FE = Failure Effort, FC = Failure Chance, FT = Failure Task Difficulty. *significance level < .01

subscales of Success Effort, Failure Ability, and Failure Effort correlated at or above .60 with the five SDQ-I subscales. Construct validity for the TSR as a measure of peer relationships was supported through correlational results. The TSR correlated at a low, but significant level with SDQ-I Peer Relations scores (r = -.23, p < .05) and failed to correlate significantly with remaining SDQ-I subscales.

Discussion The purpose of this study was to investigate socially related self-concept, attributions for social success and failure, and peer relationships for students identified as intellectually gifted and students who were high achievers at the fourth-, fifth-, and sixth-grade levels. The hypotheses that the two groups would not differ on social self-concept, social attributions, and peer relationships


NO 3

175


were based upon limited and often conf licting findings in the research literature. For the present study, the null hypothesis was generally not supported for socially related self-concept and for attributions for social success and failure. The group identified as gifted scored higher in three of four socially related self-concept areas: physical ability, physical appearance, and peer relations, as well as in the general area of self-concept. Only the area of parent relations did not result in a difference between the two groups. In addition, the group identified as gifted displayed stronger levels of attributions for social success due to internal factors of ability and effort and weaker levels of attributions for social failure due to both internal and external factors, producing an interaction effect. Results from this study add to the body of literature providing evidence that children identified as gifted are, as a group, no more vulnerable to socially related self-concept problems than the general population or, in this case, a population of high-achieving students that may have included some students potentially identifiable as gifted. Findings of this study are somewhat consistent with those of Pyryt and Mendaglio (1994), who focused on Canadian eighth and ninth graders. Findings differ, however, from Ablard’s study (1997), which also focused on eighth graders. Results based upon social attributions for success and failure differ from findings by Bell and Schindler (2002), whose fourth, fifth, and sixth graders identified as gifted did not differ in social attributions from students in regular education classes. A major difference between Bell and Schindler’s study and the present one is the make-up of the comparison groups. The heterogeneity of Bell and Schindler’s comparison group possibly contributed to the difference in findings. The differences in social attributions identified here were unexpected and compelling. For social success, gifted students attributed higher ratings to effort (M = 39.58), followed by ability (M = 35.15). This trend for the gifted group is somewhat consistent with the incremental versus entity view of ability first discussed by Dweck, as noted in Robinson (2002). Students who see their success as a function of effort more than ability ascribe to an incremental view of success, while those who see success as a function of ability more than effort ascribe to an entity view of success. Research suggests that students with an incremental view tend to be higher achievers. The difference between Success Ability and Success Effort means for the high achievers was significant and in the same direction (31.44 and 30.14, respectively) as for the 176

group identified as gifted; however, in terms of effect size, the difference could be considered negligible for the high achievers. Another theoretical perspective may offer an alternative explanation for differences in attribution results between the two groups. Some researchers have suggested that development in gifted children is asynchronous (Delisle, 1992; Silverman, 1993). Patterns of development across domains (e.g., cognitive, social, or emotional) may not necessarily follow the synchronized patterns notable in the general, nongifted population. The question arises, however, as to the effect of being identified and served as gifted on patterns of social and emotional development. Could the asynchronous patterns be due to endogenous or exogenous causes as suggested by Coleman and Cross (2000)? In other words, it is not clear if differences, when evident, are simply associated with the high cognitive and achievement levels typically associated with giftedness or with identification as gifted. Further research is needed, not only to replicate present results within and across age ranges, but also to determine possible causal relationships between attributions and identification as gifted. There were no differences between students identified as gifted and high achievers on the teacher rating of peer relationships, supporting the null hypothesis. However, there was a significant difference based on gender, with males rated higher than females by the teachers. Results for gifted and high-achieving placement groups were not consistent with those of Bell and Schindler (2002). However, ratings were based upon teacher judgments in our study and on peer judgments in the previous study. It is possible that assessment by peers, though problematic in terms of permission from parents and school authorities, constitutes a more sensitive measure, contributing to the different findings across the two studies. The difference between genders on the peer relationship ratings by teachers is ref lective of findings by Nichols (cited in Rimm, 2002), which were based upon students identified as gifted, not on high achievers. The difference between genders across both gifted and highachieving groups in the present study suggests that perhaps perceptions about academically bright boys differs from perceptions about academically bright girls as early as the late elementary grades. The children identified as gifted who participated in this study appear to be functioning in the self-concept realm at levels above their high-achieving peers and to be attributing social successes and failures in a


NO 3



pattern that promotes responsibility for interactions. However, generalizability may be limited because of the unique make-up of the gifted and high-achieving groups. The group identified as gifted was from a slightly higher socioeconomic, more professionally oriented school district, while the high-achieving group was comprised of students from this district plus students from an adjacent rural district. In addition, the high achievers likely included a few students who could potentially have been identified as gifted had they attended school in a district that actively identifies students based on state criteria. Nonetheless, these results should lead professionals providing mental health services to school-aged children to consider that those at risk for emotional or social maladjustment might be found with equal or higher probability among children not identified as gifted. Indeed, identification and placement as gifted may serve as a maintaining inf luence on positive social self-concept, particularly for children in the late elementary age group who are most often placed in pullout enrichment programs. We did not investigate grade differences in our study because of confounding effects (elementary vs. middle school placement) for fifth graders. In addition, we could not investigate causative relationships of being identified and provided special services as intellectually gifted. We recommend studies that evaluate the effects of identification as gifted, the effects of special services, and gender and grade-level effects on young children’s social self-concepts and social attributions. Children who are extremely gifted may, in fact, constitute a group of students with exceptionally different self-concepts and attributions than the individuals in our study. However, empirical evidence has not provided support for this view (Norman, Ramsay, Roberts, & Martray, 2000). Further research is warranted in this area. Children with dual classifications (gifted and another special education classification) and children identified as creatively talented, not intellectually gifted, were not studied here, providing additional areas for future research. The robust differences between identified gifted and high-achieving students on social self-concept and on social attributions are compelling. Those who worry that identification of students as gifted may have a negative social impact should be encouraged by results discussed here. These findings contribute significantly to the literature on social and emotional functioning in gifted children and suggest several areas for future research.

References Ablard, K. (1997). Self-perceptions and needs as a function of type of academic ability and gender. Roeper Review, 20, 110–115. Bain, S. K., & Reese, L. (2002). Test-retest reliability for the Student Social Attribution Scale. Unpublished manuscript. Bell, S. M., & McCallum, R. S. (1995). Development of a scale measuring student attributions and its relationship to selfconcept and social functioning. School Psychology Review, 24, 271–286. Bell, S. M., McCallum, R. S., & Doucette, J. (2003). Relationship of school-based attributions to depression. Manuscript submitted for publication. Bell, S. M., & Schindler, W. J. (2002). Collaboration at a professional development school: Investigation of variables affecting student achievement and adjustment. Tennessee Education, 31/32(1/2), 5–10. Beer, J. (1991). Depression, general anxiety, test anxiety, and rigidity of gifted junior high and high school children. Psychological Reports, 69, 1128–1130. Bracken, B. (1992). Multidimensional Self-Concept Scale: Examiner’s manual. Austin, TX: PRO-ED. Brody, L. E., & Benbow, C. P. (1986). Social and emotional adjustment of adolescents extremely talented in verbal or mathematical reasoning. Journal of Youth and Adolescence, 15, 1–18. Coie, J. D., & Dodge, K. A. (1988). Multiple sources of data on social behavior and social status in school: A cross-age comparison. Child Development, 59, 815–829. Coie, J. D., Dodge, K. A., & Coppotelli, H. (1982). Dimensions and types of social status: A cross-age perspective. Developmental Psychology, 18, 557–570. Colangelo, N., & Davis, G. A. (1991). Handbook of gifted education. Boston: Allyn and Bacon. Coleman, L. J., & Cross, T. L. (2000). Social-emotional development and the personal experience of giftedness. In K. A. Heller, F. J. Mönks, R. J. Sternberg, & R. F. Subotnik (Eds.), International handbook of giftedness and talent (pp. 203–212). Oxford, England: Elsevier Science. Coleman, J. M., & Fults, B. A. (1985). Special-class placement, level of intelligence, and the self-concepts of gifted children: A social comparison perspective. Remedial and Special Education, 6, 7–11. CTB/McGraw-Hill (1996). Comprehensive test of basic skills. Monterrey, CA: CTB/McGraw-Hill. Delisle, J. (1992). Guiding the social and emotional development of gifted youth: A practical guide for educators and counselors. New York: Longman. Frentz, C., Gresham, F. M., & Elliott, S. N. (1991). Popular, controversial, neglected, and rejected adolescents: Contrasts of social competence and achievement differences. Journal of School Psychology, 29, 109–120.


NO 3

177

S O C I A L / E M OT I O N A L S TAT U S O F G I F T E D C H I L D R E N Gough, H. G., & Heilbrun, A. B. (1983). The Adjective Checklist manual. Palo Alto, CA: Consulting Psychologists Press. Grossberg, I., & Cornell, D. (1988). Relationship between personality adjustment and high intelligence: Terman versus Hollingsworth. Exceptional Children, 55, 266–272. Hoge, R. D., & Renzulli, J. S. (1993). Exploring the link between giftedness and self-concept. Review of Educational Research, 63, 449–465. Jackson, L. D., & Bracken, B. A. (1998). Relationship between students’ social status and global-domain specific self-concepts. Journal of School Psychology, 36, 233–246. Kennedy, J. (1990). Determinants of peer social status: Contributions of physical appearance, reputation, and behavior. Journal of Youth and Adolescence, 19, 233–244. Kulik, J. A., & Kulik, C. L. (1992). Meta-analytic findings on grouping programs. Gifted Child Quarterly, 36, 73–77. Lajoie, S. P., & Shore, B. M. (1981). Three myths? The overrepresentation of the gifted among dropouts, delinquents, and suicides. Gifted Child Quarterly, 25, 138–143. Lombroso, C. (1895). The man of genius. New York: Scribner. Marsh, H. W. (1987). The big-fish-little-pond effect on academic self-concept. Journal of Educational Psychology, 79, 280–295. Marsh, H. W. (1988). Self-description questionnaire–I. San Antonio, TX: The Psychological Corporation. Marsh, H. W., & Shavelson, R. (1985). Self-concept: Its multifaceted, hierarchical structure. Educational Psychologist, 20, 107–123. Nail, J. M., & Evans, J. G. (1997). The emotional adjustment of gifted adolescents: A view of global functioning. Roeper Review, 20, 18–21. Neihart, M. (1999). The impact of giftedness on psychological well-being: What does the empirical literature say? Roeper Review, 22, 10–17. Neihart, M., Reis, N. M., Robinson, S. M., & Moon, S. M. (Eds.). (2002). The social and emotional development of gifted children: What do we know? Waco, TX: Prufrock Press. Norman, A. D., Ramsay, S. G., Roberts, J. L. & Martray, C. R. (2000). Effects of social setting, self-concept, and relative age on the social status of moderately and highly gifted students. Roeper Review, 23, 34–39.

178

Pyryt, M. C., & Mendaglio, S. (1994). The multidimensional self-concept: A comparison of gifted and average-ability adolescents. Journal for the Education of the Gifted, 17, 299–305. Reis, S. M. (2002). Gifted females in elementary and secondary school. In M. Neihart, S. M. Reis, N. M. Robinson, & S. M. Moon (Eds.), The social and emotional development of gifted children: What do we know? (pp. 125–135). Waco, TX: Prufrock Press. Rimm, S. (2002). Peer pressures and social acceptance of gifted students. In M. Neihart, S. M. Reis, N. M. Robinson, & S. M. Moon (Eds.), The social and emotional development of gifted children: What do we know? (pp. 13–18). Waco, TX: Prufrock Press. Robinson, N. M. (2002). Individual differences in gifted students’ attributions for academic performance. In M. Neihart, S. M. Reis, N.M. Robinson, & S.M. Moon (Eds.), The social and emotional development of gifted children: What do we know? (pp. 61–69). Waco, TX: Prufrock Press. Silverman, L. (1993). Counseling the gifted and talented. Denver: Love. Solano, C. H. (1987). Stereotypes of social isolation and early burnout in the gifted: Do they still exist? Journal of Youth and Adolescence, 16, 527–539. Terman, L. M., & Oden, M. H. (1947). The gifted child grows up: Twenty-five years’ follow-up of a superior group. Stanford, CA: Stanford University Press.

Author Note We would like to acknowledge and thank Shawn J. Bourgeois of Jefferson Parish Schools (Louisiana) for his help in preparing the literature review section of this manuscript. Address correspondence concerning this article to Sherry K. Bain, Department of Educational Psychology and Counseling, University of Tennessee, 525 Claxton Complex, Knoxville, TN, 37996-3452.


NO 3
