Self-schemas, anxiety, somatic and depressive symptoms in socially withdrawn children and adolescents.

Abstract. The present study examined the hypothesis that negative self-schemas and a range of internalizing problems may be associated with social withdrawal in late childhood and early adolescence. Using in-school peer nominations from the Revised Class Play, 132 eleven- to thirteen-year-olds were classified as withdrawn (n = 40; 22 girls, 18 boys), aggressive (n = 41, 21 girls, 20 boys), or normative control (n = 51; 28 girls, 23 boys). Children completed a self-referent rating task (SRRT), as well as the Youth Self-Report (YSR). The groups were compared in terms of their YSR scores for internalizing problem subtypes, as well as their endorsement and recall of positive and negative descriptors from the SRRT Compared to the other groups, socially withdrawn children/adolescents reported higher levels of each internalizing problem--that is, shy/withdrawn behaviors, anxiety, depressive symptoms, and somatic problems. Furthermore, although differences in recall for SRRT positive and negative descriptors were not found, socially withdrawn children endorsed fewer positive descriptors and more negative descriptors than did the normative control and aggressive groups. Moreover, fewer positive self-perceptions were a stronger discriminator between the groups than were more negative self perceptions. These findings were interpreted within a risk/protective framework of adjustment.

**********

Children's relationships with their peers play an important role in many areas of psychological development (Asher & Coie, 1990; Bukowski, Newcomb, & Hartup, 1996; Piaget, 1932; Rubin, Bukowski, & Parker, 1998; Sullivan, 1953). Children who interact infrequently with their peers may lack the opportunity to develop the social skills necessary for the development of healthy interpersonal relationships (e.g., Hartup, 1983). Such social isolation may be the consequence of overt rejection by peers, as a result of aggressive or aversive behavior; however, it also may be a consequence of the child's own tendency to remove him/herself from social interaction, either because of social anxiety or because of a preference for solitude (Rubin, Burgess, & Coplan, 2002; Rubin & Mills, 1988; Younger & Daniels, 1992). Various labels have been used to describe this latter self-imposed social isolation, including social withdrawal, shyness, and behavioral inhibition (see Rubin & Asendorpf, 1993; Rubin & Burgess, 2001).

For several reasons, socially withdrawn children have been studied less than children who are aggressive and disruptive. In the school environment, shy/withdrawn behaviors are less problematic than aggressive behaviors, causing neither disruption in the classroom nor offense to peers. Furthermore, in contrast to concerns about aggressive children, researchers have not considered socially withdrawn children to be at risk for later psychological or emotional difficulties. Rubin and colleagues (Rubin, Burgess, Kennedy, & Stewart, 2003), however, have proposed a developmental pathway model that illustrates the correlates and potential outcomes of social withdrawal at various stages of development. According to this model, early childhood shyness/withdrawal may eventually lead to psychological disorders of an internalizing nature, particularly anxiety and depression.

Yet few studies have actually assessed the risk status of socially withdrawn children. Limited data exist concerning the concurrent psychological adjustment of socially withdrawn children. Some studies suggest that, in early childhood, withdrawn behavior in its solitary-constructive/exploratory or unsociable form is not associated with intrapersonal or interpersonal problems (Coplan, Rubin, Fox, Calkins, & Stewart, 1994; Ladd & Burgess, 1999) and is not unusual (Rubin & Clark, 1983). With increasing age, however, social withdrawal may become more problematic. Social withdrawal in mid- to late-childhood has been associated with negative self-perceptions of social competence (e.g., Hymel, Bowker, & Woody, 1993; Rubin, Hymel, & Mills, 1989), as well as with rejection, loneliness, victimization, and depressed mood (e.g., Bell-Dolan, Reaven, & Peterson, 1993; Boivin, Hymel, & Bukowski, 1995; Hymel, Rubin, Rowden, & LeMare, 1990; Rubin et al., 1989). Into adolescence, problems with felt security, low self-perceived social competence, and loneliness become evident (Rubin, Chen, McDougall, Bowker, & McKinnon, 1995). Thus, withdrawn children who experience social anxiety, low social interaction, and negative self-perceptions may subsequently develop disorders of an internalizing nature.

With these psychosocial and behavioral difficulties, it is surprising that studies on shy, socially withdrawn children typically have not been recognized for their relevance and value to the field of education. Social withdrawal usually is evident in the classroom on a regular basis (Rubin, Burgess, & Coplan, 2002; Rubin, Burgess, Kennedy, & Stewart, 2003), even once familiarity is developed with peers and teachers. These children consistently interact and communicate less, thereby interfering with or negatively affecting their educational process and learning potential. Children and adolescents who are very shy or socially anxious do not talk much, speak quietly, and rarely initiate conversations with peers or teachers (Burgess, 2003; Younger, Schneider, Wadeson, Guirguis, & Bergeron, 2000). Additionally, they are nervous about participating in classroom or group discussions because they are worried about how they will be perceived if they do not ask a "good" question or if they get an answer wrong.

Such reticence to participate in classroom discussions, answer questions, ask for-help when needed, and be their own advocates (particularly near and during the adolescent period) lower these students' involvement and engagement in their own learning--factors that directly affect academic achievement, performance, and liking school or wanting to go to school (Ladd, Birch, & Buhs, 1999). Therefore, studying withdrawn children's psychological adjustment has great importance for educational outcomes.

Cognitive models of self-schemas and information-processing provide a theoretical framework through which to examine a possible link between social withdrawal and psychological maladjustment. Negative self-schemas, which entail endorsing negative information about oneself and recalling such information, long have been considered central to the development of depression and anxiety disorders (e.g., Beck, 1967/1972; Cicchetti & Schneider-Rosen, 1986). Indeed, several studies have shown that negative self-schemas are associated with depressive affect in adults (Derry & Kuiper, 1981; Kuiper & Derry, 1982; Kuiper, Olinger, MacDonald, & Shaw, 1985), as well as in children (Zupan, Hammen, & Jaenicke, 1987). Consequently, children showing signs of negative self-schemas may be at risk for internalizing disorders, such as depression.

Currently, no data exist regarding such self-schemas and the processing of personal information in withdrawn children. Nevertheless, several studies have supported the applicability of the self-schema model to depression in children. Compared to nondepressed children, depressed children have been found to show poorer recall for positive self-descriptive words (Hammen & Zupan, 1984; Zupan et al., 1987), view themselves more negatively (Zupan et al., 1987), and voice more negative thoughts about themselves (Sanders, Dadds, Johnston, & Cash, 1992). Indeed, Jaenicke et al. (1987) reported that children who have mothers with major affective disorder had a more negative self-concept and recalled fewer positive self-descriptives than did those with nondepressed mothers. Thus, the capacity for cognitive bias and selective recall exists in children as well as adults; moreover, children at risk for depression display negative cognitions about the self. Such research, however, has been conducted with depressed rather than withdrawn children.

The information-processing methodology offers a measure that may be suitable for testing hypotheses about self-schemas in withdrawn children. If socially withdrawn children are at risk for internalizing disorders, then one factor contributing to their experience of such distress might be the presence of negative self-schemas and/or the absence of positive self-schemas.

Thus, the first objective of this research was to assess the positive and negative self-schemas of socially withdrawn children. To accomplish this goal, children were presented with a series of positive and negative descriptors from multiple domains (i.e., behavioral, social, emotional, physical appearance) to reflect the multidimensional nature of the self-concept. They were asked to indicate whether or not each word was self-descriptive (cf. Hammen & Zupan, 1984); following the presentation of the entire series, their incidental recall for these descriptors was assessed. We hypothesized that withdrawn children would be likely to endorse and recall more negative and fewer positive descriptors than would aggressive children and normative control children.

The second objective was to determine whether withdrawn children experience internalizing problems in early adolescence and whether possible internalizing problems extend beyond depressive symptoms. The studies reviewed indicate that withdrawn children experience loneliness in middle to late childhood and may be prone to depressed mood per se; but only the Children's Depression Inventory (Kovacs, 1982) has been utilized. Researchers have yet to employ the full range of internalizing symptoms using self-report for somatic complaints, shy/withdrawn behavior, anxiety, and depression. It was anticipated that an older group of withdrawn children (preadolescent to adolescent age) might experience anxiety and depression. We also speculated that they might report somatic symptoms, which would suggest that their social-behavioral maladjustment may be associated with physical expressions of discomfort. Furthermore, we wondered whether they perceive themselves as shy and withdrawn. It does not appear that investigators have yet conducted such self-assessments among older children and adolescents. Self-reports at this age have better validity than other sources, such as teachers and parents, for the assessment of internalizing problems (e.g., Bell-Dolan et al., 1993).

Finally, a third objective was to compare the two major forms of childhood social maladjustment: social withdrawal and aggression. Children who exhibit overt (direct, confrontive) forms of aggression, as opposed to covert (indirect, relational) forms, were targeted because of the short-term and long-term evidence indicating that this type of aggression poses risk for psychopathology. By comparing withdrawn and aggressive children within the same study, it was possible to determine whether different behavioral interaction styles, manifest in late childhood and early adolescence, are linked to unique self-concept or behavioral-emotional profiles, or whether they are associated with similar profiles in these domains. In other words, does the nature of withdrawn children's psychological functioning relate to their particular characteristics, or is it a reflection of childhood maladjustment in general?

To address the above objectives, we examined the psychological adjustment of socially withdrawn and aggressive children in 6th and 7th grades, assessing their self-perceptions and behavioral-emotional functioning on two measures: a self-referent rating task and the Youth Self-Report.

The Self Referent Rating Task (SRRT) was developed to measure children's endorsement and recall of positive and negative self-descriptions. Children were presented with a series of positive and negative adjectives and asked to indicate whether each item described them or not. This procedure was followed by an incidental recall task, during which their recall of positive and negative descriptors was assessed. It was predicted that withdrawn children would endorse and recall more negative descriptors and fewer positive descriptors compared to control and aggressive children.

The Youth Self-Report (Achenbach, 1991) measure yields a clinically based assessment of behavioral and emotional problems. Our specific focus was on the Internalizing scale, which includes anxiety, depression, shy-withdrawn behavior, and somatic complaints. It was hypothesized that socially withdrawn children would have higher internalizing scores on the behavioral and emotional adjustment measure than their aggressive and average peers. By contrast, we expected that aggressive children would have higher externalizing scores, as found in previous work.

Method

Participants

The sample was recruited from 15 sixth- and seventh-grade classes in five elementary schools in a region that ranged from suburban to moderately urban in size and location. Income level was estimated to be low-middle to middle, based on what was typical for the geographical area.

Parents' informed consent was obtained in the fall of the school year, and 80 percent or higher permission rates were achieved in all 15 classrooms. In the fall and winter of 6th and 7th grades, three groups (n = 132) were selected from the larger normative sample (N = 281) (selection criteria described below). The composition of the sample was predominantly Caucasian.

Through group administration, peer nominations were collected from the Revised Class Play, and then participants were identified as either withdrawn (n = 40; 18 boys, 22 girls), aggressive (n = 41; 20 boys, 21 girls), or normative control (n = 51, 23 boys, 28 girls).

Identification of Risk and Control Groups Peer ratings of children's aggressive and withdrawn behaviors were obtained from the Aggressive-Disruptive and Sensitivity/ Isolation factors of the Revised Class Play (RCP; Masten, Morison, & Pellegrini, 1985). The RCP is a peer nomination measure that consists of 30 items; it yields three reliable independent factors, labeled Sensitivity-Isolation, Aggression-Disruption, and Sociability-Leadership.

The RCP asks children to nominate peers for roles in an imaginary class play depicting sociable, aggressive, and withdrawn behaviors. During group administration, children were requested to nominate up to three male and three female classmates who best fit each of the 30 behavioral descriptors. Only children with parental consent to participate could either nominate peers or be nominated. Nominations received from peers were used to compute factor scores for each child following procedures outlined by Masten et al. (1985).

Passive Withdrawal Scale. Rubin and Mills (1988) proposed that the Sensitivity-Isolation factor may confound "passive" withdrawal from the peer group with active rejection by the peer group, and recommended using a subset of the Sensitivity-Isolation items to assess passive-withdrawal, a suggestion that was empirically supported by Younger and Daniels (1992). Thus, although the entire Sensitivity-Isolation scale was administered, passively withdrawn children were identified using the following three items only: "Someone who is very shy," "Someone whose feelings get hurt easily," and "Someone who would rather play alone than with others."

Aggressive-Disruptive Scale. The Aggressive-Disruptive scale includes eight items, such as "A person who gets into a lot of fights," "Somebody who teases other children too much," and "Someone who interrupts when other children are speaking."

For both scales, the number of nominations received by each child for each item was tallied and then standardized within sex and classes to permit appropriate comparisons. These standardized scores were then summed to compute scores for the passive withdrawal and aggressive-disruptive factors. Standardizations adjusted for variations in class size, number per grade, and number of boys and girls. Higher scores meant stronger peer perceptions of the identified behavior.

Inclusion criteria were as follows: Forty children who scored above the 80th percentile on the passive withdrawal index (Rubin & Mills, 1988) of the RCP Isolation-Sensitivity scale and below the 50th percentile of the present sample on the Aggression scale were identified as the passively withdrawn group. Forty-one children who scored above the 80th percentile on the RCP Aggression-Disruptive scale and below the 50th percentile on the passive withdrawal subscale were classified as the aggressive group. Fifty-one children who scored below the 50th percentile on the passive withdrawal subscale and below the 50th percentile on the Aggression scale were selected for the normative control (nonwithdrawn/nonaggressive) group.

Measures of Psychological Adjustment

Self-Schema Task. The Self-Referent Rating Task (SRRT), which was developed from Hammen and Zupan's (1984) study as well as our own pilot study, was employed to measure children's endorsement and recall of positive and negative words as self-descriptive. The goals of the pilot study were twofold: first, to ensure that the youngest participants would comprehend all of the words; and second, to have the words rated for positive versus negative content. Twelve 11-year-old children (7 boys and 5 girls) served in the pilot study and were presented with 51 words, which then were assessed for both comprehension and desirability. Children were asked to define each word and then categorize it as positive, negative, or neutral. Only those words found to be easily comprehended by all 12 participants, and for which there was greater than 80 percent agreement on their positive or negative content, were retained. A total of 18 positive words (e.g., "nice," "smart," "popular") and 18 negative words (e.g., "bad," "ugly," "boring") remained. Table 1 presents these positive and negative descriptors. As expected, given the contents of the lists of descriptors, we found a negative correlation between the number of positive and negative items children in our sample endorsed as self-descriptive (r = -.39, p < .01).

Behavioral-Emotional Adjustment. The Youth Self-Report (YSR; Achenbach, 1991) is a ll2-item self-report measure of behaviors and emotions suitable for children/adolescents between 11 and 18 years old. Through individual administration, students completed this measure in booklet form and were debriefed afterward. The YSR identifies syndromes of problems that fall under the two broad groupings of Internalizing and Externalizing factors. Subjects chose one of three responses to each item (not true, sometimes true, often true), describing the way he or she felt within the past six months. Previous research on the YSR has shown test-retest reliabilities for the Externalizing scale of .76 for a one-week interval and .49 for a seven-month interval, and .67 and .52 for the Internalizing scale, for the one-week and seven-month intervals, respectively (Achenbach, 1991). There is evidence for content validity, as most YSR items discriminate between clinical and normal samples. There is also evidence for criterion-related validity of the YSR scales, as the scale scores discriminate between referred and non-referred youths. The YSR Internalizing scale consists of the following three subscales: anxious/depressive symptoms, somatic complaints, and withdrawn behavior. In our sample, these three subscales were found to correlate .92, .72, and .77, respectively, with the Internalizing scale.

Procedure

Both phases of the study were conducted at school with all children for whom consent was obtained. A senior doctoral student in psychology conducted the interviews.

Phase 1. The first phase of the study involved group administration of the peer nominations on the RCP. Children were instructed to choose from a class list those classmates who best fit each statement. Nominations pertaining to boys were conducted separately from nominations for girls, because girls are more frequently nominated for indices of withdrawal (Younger & Daniels, 1992) and both sexes were required for the sample.

In order to form the groups, RCP data analysis was conducted on the whole sample of 281 children. Raw scores for each child were tallied by adding the number of nominations received from classmates for all RCP items, divided by the number of raters in the class. Total scores then were computed for each scale. Before converting data to z-scores, a log transformation was used to reduce skewness (Tabachnick & Fidell, 1989).

Phase 2. During the second phase, the dependent measures were administered to children who met the inclusion criteria. Each child was tested individually and completed the dependent measures at one point in time. The SRRT task involves both a self-endorsement phase and a surprise recall phase. In order not to contaminate this information-processing task with other stimuli, the administration of the SRRT always preceded that of the YSR.

The items making up the SRRT (see Table 1) were presented serially, and participants were asked to indicate whether the adjective described themselves or not ("Is this like you?"; circle Yes or No). All words were displayed on 3 x 5 index cards contained in a booklet, and word order was randomized for each subject. Immediately following the self-referent rating task, participants were unexpectedly asked to recall as many of the words as possible. During a four-minute maximum period, the experimenter recorded the number of positive and negative self-descriptions recalled. After the SRRT administration, the children completed the inventory of internalizing and externalizing behaviors (YSR) presented to them in booklet format. The total assessment time was approximately 30 to 45 minutes per child.

Results

A between-groups differential design was utilized to test the hypotheses. A between-subjects multivariate analysis of variance (MANOVA) was conducted to assess differences between socially withdrawn, aggressive, and normative control groups on the following dependent variables: number of positive words endorsed as self-descriptive (positive self-perceptions), number of negative words endorsed as self-descriptive (negative self-perceptions), number of negative words recalled, number of positive words recalled, internalizing scores, and externalizing scores.

Preliminary MANOVAs indicated no significant main or interaction effects of sex or grade. Consequently, these variables were not included as factors in the analysis.

The overall MANOVA results, using Pillais trace test statistic as the criterion (Pillais value = .58), indicated a significant multivariate main effect of Group, F (12, 250) = 8.46, p < .001. No significant Group by Sex interactions emerged. As the omnibus MANOVA showed a statistically significant effect, the univariate results were examined to determine where the differences between the three groups occurred.

Self-Referent Rating Task (SRRT)

We predicted that socially withdrawn children would endorse a higher number of negative words and a lower number of positive words as being self-descriptive. The analyses below reveal that this prediction was supported.

Positive Self-Perceptions. Univariate analysis (ANOVA) indicated a significant effect of Group on positive self-perceptions, F (2, 129) = 20.93, p < .001. Tukey-HSD post hoc comparisons revealed that the withdrawn group endorsed significantly fewer positive words as self-descriptive than did the control and aggressive groups (p < .001), which did not significantly differ from each other in their endorsement of positive self-descriptors. Table 3 contains the summary descriptive statistics for the self-perception variables according to group.

Negative Self-Perceptions. A significant effect of Group on negative self-perceptions also was found, F(2, 129) = 12.21, p < .001. Post hoc Tukey-HSD tests showed that the withdrawn group endorsed a significantly higher number of negative words as self-descriptive than did the control and aggressive groups (p < .001 and p < .01, respectively), which did not differ from one another (see Table 3).

To discern whether one kind of self-perception was a stronger discriminator between groups than the other (i.e., positive versus negative), a discriminant function analysis showed that the positive self-perception variable accounted for most of the variance, or was most strongly associated with group differences. In the first function, the standardized canonical discriminant function coefficients were .69 for positive self-perception (Yes to positive words), -.40 for negative self-perception (Yes to negative words), and -.24 for the internalizing T-score. Thus, in this sample, the best variable that discriminates between the groups (i.e., the strongest contributor) is positive self-perceptions.

Incidental Recall Task. We predicted that withdrawn children would recall a smaller number of positive and a greater number of negative words than the control or aggressive children. However, no significant differences emerged between the three groups in terms of recall of either positive words, F(2, 129) = 1.34, p > .05, or negative words, F(2, 129) = .80, p > .05.

Youth Self-Report

We predicted that the groups would differ in their behavioral-emotional functioning, represented by the dependent variables of internalizing and externalizing scores on the YSR. Specifically, it was expected that socially withdrawn children would have higher internalizing scores than normative control and aggressive children. Univariate F-tests revealed a significant effect of Group on internalizing scores using Bonferroni correction, F(2, 129) = 9.81),p < .001. Tukey tests revealed that the withdrawn group had higher Internalizing T-scores than the control and aggressive groups, which did not differ significantly from each other (see Table 3).

Having obtained a significant difference between the socially withdrawn and normative control groups on the Internalizing scale, a second MANOVA was run to examine the hypothesized differences between these two groups on the three YSR internalizing subscales. The overall MANOVA was significant (F (3, 87) = 9.83, p < .001), and, as anticipated, univariate analyses revealed a significant difference between the withdrawn and normative control groups in terms of anxious/depressive symptoms (F (1, 89) = 20.02, p < .001), somatic complaints (F (1, 89) = 4.70, p < .05), and shy/withdrawn behavior (F (1, 89) = 24.30, p < .001). Table 4 contains the summary descriptive statistics for each of these internalizing variables according to group.

Lastly, it was predicted that the aggressive group would have higher externalizing scores than would control or withdrawn children. For this analysis, the univariate F was also significant using Bonferroni correction, F (2, 129) = 12.36, p < .001. Tukey tests indicated that the aggressive group had significantly higher externalizing scores than the withdrawn group and the control group, which did not differ from each other (see Table 3).

Discussion

This study sought to determine whether socially withdrawn children differed from normative control children and aggressive children with respect to self-schemas and behavioral-emotional functioning, as assessed in school settings. It is important to note that our sample of children and young adolescents was drawn from a community rather than a clinic-referred population, and our investigation was conducted in the natural environment of school. Our hypotheses about self-schemas were partially confirmed. As predicted, withdrawn children endorsed fewer positive and more negative words as self-descriptive than did aggressive or control children. These findings are consistent with previous research with younger children. For example, Rubin, Hymel, and colleagues found lower self-perceived social competence among withdrawn children of elementary school age (Hymel et al., 1990; Rubin et al., 1989; Rubin, Chen, & Hymel, 1993). The current study illustrates that an older sample (i.e., late childhood to early adolescence) of withdrawn children had a poor self-concept in multiple domains, and not only in a social domain.

Notably, the findings of the present study go beyond previous work by finding that socially withdrawn children endorsed fewer positive words as self-descriptive. The fact that shy/withdrawn children displayed relatively fewer positive self-perceptions dovetails well with the literature comparing normal versus depressed individuals. Abramson and Alloy (1990) have noted that normal, nondepressed adults show positive biases in their cognitions about the self. This optimistic, as opposed to realistic, manner of perceiving oneself is thought to be important for adaptive reasons. The withdrawn children in this study, however, appeared to show less of such a positive bias to protect themselves than did the non-withdrawn children and aggressive children. It might, therefore, be more difficult for them to maintain a positive or non-depressed state when faced with social or emotional adversity.

Although withdrawn children displayed both a greater level of negative self-perceptions and lower level positive self-perceptions than the other groups, their lower positive self-evaluations explained more of the variance than did their negative self-evaluations. There may be important practical implications of this finding for those working with socially withdrawn children and adolescents. It seems that having fewer positive self-perceptions may be more important to consider than the existence of negative self-perceptions.

It is interesting that whereas withdrawn children acknowledged significantly less positive and more negative self-perceptions than did normative control and aggressive children, aggressive children did not significantly differ from control children with respect to either type of self-perception. Aggressive children viewed themselves just as positively and no more negatively than did control children. This finding is congruent with other studies that have found aggressive children overestimate and misinterpret their own competencies relative to peers' evaluations of them (e.g., Hymel, Bowker, & Woody, 1993). Although aggressive children appeared satisfied with themselves, they did acknowledge externalizing behaviors on the clinical measure. Moreover, aggressive children were similar to controls on positive self-perceptions: possibly they lack the self-awareness of having fewer desirable traits than controls, or they do not have relatively fewer positive characteristics compared to controls, and they feel good about themselves in terms of positive characteristics.

Children who are shy, overly sensitive, and interact infrequently with peers tend to experience a relatively poor self-concept. Furthermore, the findings that the withdrawn children, but not the aggressive children, differed significantly from normative control children with respect to self-perceptions suggest that such maladaptive self-perceptions may be uniquely important to social withdrawal as compared to other forms of maladjustment. During late childhood to early adolescence, then, withdrawn individuals have negative cognitions about themselves relative to their peers. Perhaps such negative cognitions, which are also a core feature of depression and anxiety, place socially withdrawn children at further risk for developing such internalizing disorders.

Contrary to expectations, the groups did not differ on recall of positive or negative descriptors. Thus, this second aspect of self-schema function was not supported in the current findings. Although puzzling, especially in light of our findings concerning self-evaluations, the lack of bias in socially withdrawn children's recall may reflect the fact that the withdrawn children in this sample were not clinically depressed. Indeed, research with adults suggests that recall of negative personal information may be mood dependent (Hammen & Zupan, 1984), with current depressed mood being the strongest predictor of negative self-schema function (Kuiper et al., 1985; Zupan et al., 1987).

In terms of behavioral-emotional functioning on the YSR, a good indication of risk status in late childhood and early adolescence was gleaned. As predicted, socially withdrawn children/adolescents reported that they experienced more problems of an internalizing kind than did normative control children and aggressive children. These findings indicate that children who are shy, who isolate themselves, and who interact infrequently with peers tend to experience internalizing problems. Specifically, this sample of withdrawn children/adolescents, drawn from the general school population, had higher levels of anxiety symptoms, depressive symptoms, and somatic complaints than did other children. Moreover, they reported themselves as shy and as displaying withdrawn behaviors more often than others. Unlike aggressive children, withdrawn children's level of self-awareness is fairly accurate. This higher level of psychological insight might give them relatively better ability to benefit from intervention. Particularly noteworthy is the finding of somatic complaints in the withdrawn group. The fact that these children from a nonclinical, classroom sample are experiencing physical symptoms with no known medical cause (e.g., headaches, stomach aches, dizziness, fatigue, nausea) on a regular basis at school is rather disturbing.

Even though the design was not longitudinal and no conclusions can be drawn about the outcomes of withdrawal, these findings do fit with the notion that these children may be at risk for internalizing disorders in adolescence and adulthood, as proposed by Rubin, Burgess, Kennedy, and Stewart (2003). Social withdrawal in preadolescence was indeed related to internalizing problems in our study. Longitudinal studies following such children into later adolescence and adulthood might elucidate causal mechanisms or mediators between social withdrawal and internalizing behaviors or emotions. As withdrawn children already experience interpersonal difficulties, according to themselves and their peers, they might be at continued risk for social anxiety and relationship problems (see Burgess, Wojslawowicz, Rubin, Rose-Krasnor, & Booth-LaForce, 2006).

Implications for Practice

Having the knowledge that shy, withdrawn children are prone to negative affectivity could be useful for tailoring individual and group treatment programs to meet their specific needs. Moreover, school counselors and teachers who are on the "front lines" could offer help during school hours, as currently done with lunch/recess time groups for "making friends" and "children of divorce." Currently, these children do not receive much attention and are seldom targeted for groups at school or referrals for evaluation and treatment, nor are they necessarily recognized as being at-risk for psychological distress or lack of involvement and participation in their own learning. In contrast to aggressive children, who are often referred for treatment by others because of such disruptive and aversive behaviors as verbal threats or physical fights, shy/withdrawn children are frequently overlooked because their inoffensive behaviors seldom impinge on others in the school environment.

Future interventions, either in the form of individual treatment or group social skills programs, should incorporate recent knowledge about the self-perceptions and the emotional/behavioral functioning of these underserved children. Better still might be a preventive approach whereby these individuals are identified in early childhood at school and elsewhere, then given the appropriate research and clinical attention.

References

Abramson, L. Y., & Alloy, L. B. (1990). Search for the "negative cognition" subtype of depression. In C. D. McCann & N. S. Endler (Eds.), Depression: New directions in theory, research, and practice (pp. 77-109). Toronto, ON: Wall & Emerson.

Achenbach, T. M. (1991). Manual for the Youth Self-Report and 1991 profile. Burlington, VT: University of Vermont Department of Psychiatry.

Asher, S. R., & Coie, J. D. (1990). Peer rejection in childhood. New York: Cambridge University Press.

Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press. (original work published 1967)

Bell-Dolan, D., Reaven, N. M., & Peterson, L. (1993). Depression and social functioning: A multidimensional study of the linkages. Journal of Clinical Child Psychology, 22, 306-315.

Boivin, M., Hymel, S., & Bukowski, W. (1995). The roles of social withdrawal, peer rejection, and victimization by peers in predicting loneliness and depressed mood in childhood. Development and Psychopathology, 7, 765-785.

Bukowski, W. M., Newcomb, A. F., & Hartup, W. W. (1996). The company they keep: Friendship in childhood and adolescence. New York: Cambridge University Press.

Burgess, K. B., Wojslawowicz, J. C., Rubin, K. H., Rose-Krasnor, L., & Booth-LaForce, C. (2006). Social information processing and coping strategies of shy/withdrawn and aggressive children: Does friendship matter? Child Development, 77, 371-383.

Burgess, K. B., Wojslawowicz, J. C., Rubin, K. H., Rose-Krasnor, L., & Booth, C. L. (2003, April). The "Extended Class Play": A longitudinal study of its factor structure, reliability, and validity. Presented at the biennial meeting of the Society for Research in Child Development, Tampa, FL.

Cicchetti, D. & Schneider-Rosen, K. (1986). An organizational approach to childhood depression. In M. Rutter, C. Izard, & P. Read (Eds.), Depression in young people: Developmental and clinical perspectives (pp. 71-134). New York: Guilford.

Coplan, R. J., Rubin, K. H., Fox, N. A., Calkins, S. D., & Stewart, S. L. (1994). Being alone, playing alone, and acting alone: Distinguishing among reticence and passive and active solitude in young children. Child Development, 65, 129-137.

Derry, P. A., & Kuiper, N.A. (1981). Schematic processing and self-reference in depression. Journal of Abnormal Psychology, 90, 286297.

Hammen, C., & Zupan, B. A. (1984). Self-schemas, depression, and the processing of personal information in children. Journal of Experimental Child Psychology, 37, 598-608.

Hartup, W. W. (1983). Peer relations. In P. H. Mussen (Series Ed.) & E. M. Hetherington (Vol. Ed.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (4th ed.). New York: Wiley.

Hymel, S., Bowker, A., & Woody, E. (1993). Aggressive versus withdrawn unpopular children: Variations in peer and self-perceptions in multiple domains. Child Development, 64, 879-896.

Hymel, S., Rubin, K. H., Rowden, L., & LeMare, L. (1990). Children's peer relationships: Longitudinal prediction of internalizing and externalizing problems from middle to late childhood. Child Development, 61, 2004-2021.

Jaenicke, C., Hammen, C., Zupan, B., Hiroto, D., Gordon, D., Adrian, C., & Burge, D. (1987). Cognitive vulnerability in children at risk for depression. Journal of Abnormal Child Psychology, 15, 559-572.

Kovacs, M. (1982). Children's depression inventory. Pittsburgh, PA: Western Psychiatric Institute.

Kuiper, N. A. & Derry, P.A. (1982). Depressed and non-depressed content self-reference in mild depressives. Journal of Personality, 50, 67-80.

Kuiper, N. A., Olinger, L. J., MacDonald, M. R., & Shaw, B. F. (1985). Self-schema processing of depressed and non-depressed content: The effects of vulnerability to depression. Social Cognition, 3, 77-93.

Ladd, G. W., Birch, S., & Buhs, E. (1999). Children's social and scholastic lives in kindergarten: Related spheres of influence? Child Development, 70, 1373-1400.

Ladd, G. W., & Burgess, K. B. (1999). Charting the relationship trajectories of aggressive, withdrawn, and aggressive/withdrawn children during early grade school. Child Development, 70, 910-929.

Masten, A. S., Morison, P., & Pellegrini, D. S. (1985). A revised class play method of peer assessment. Developmental Psychology, 21, 523-533.

Piaget, J. (1932). The moral judgment of the child. London: Routledge and Kegan Paul.

Rubin, K. H., & Asendorpf, J. B. (1993). Social withdrawal, inhibition, and shyness in childhood. Hillsdale, NJ: Erlbaum.

Rubin, K. H., Bukowski, W., & Parker, J. G. (1998). Peer interactions, relationships, and groups. In W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (5th ed., pp. 619-700). New York: Wiley.

Rubin, K. H., & Burgess, K. B. (2001). Social withdrawal and anxiety. In M. W. Vasey & M. R. Dadds (Eds.), The developmental psychopathology of anxiety (pp. 407-434). New York: Oxford University Press.

Rubin, K. H., Burgess, K. B., & Coplan, R. J. (2002). Social withdrawal and shyness. In P. K. Smith & C. H. Hart (Eds.), Handbook of social development (pp. 329-352). London, UK: Blackwell.

Rubin, K. H., Burgess, K. B., Kennedy, A. E., Stewart, S. (2003). Social withdrawal in childhood. In E. J. Mash & R. A. Barkley (Eds.), Child psychopathology (2nd ed., pp. 372-406). New York: Guilford Publications.

Rubin, K. H., Chen, X., & Hymel, S. (1993). Socio-emotional characteristics of extremely withdrawn and extremely aggressive children. Merrill-Palmer Quarterly, 39, 518-534.

Rubin, K. H., Chen, X., McDougall, P., Bowker, A., & McKinnon, J. (1995). The Waterloo Longitudinal Project: Predicting internalizing and externalizing problems in adolescence. Development and Psychopathology, 7, 751-764.

Rubin, K. H., & Clark, M. L. (1983). Preschool teachers' ratings of behavioral problems: Observational, sociometric, and social-cognitive correlates. Journal of Abnormal Child Psychology, 11, 273-285.

Rubin, K. H., Hymel, S., & Mills, R. S. (1989). Sociability and social withdrawal in childhood: Stability and outcomes. Journal of Personality, 57, 237-255.

Rubin, K. H., & Mills, R. S. L. (1988). The many faces of social isolation in childhood. Journal of Consulting and Clinical Psychology, 56, 916-924.

Sanders, M. R., Dadds, M. R., Johnston, B. M., & Cash, R. (1992). Childhood depression and conduct disorder: I. Behavioral, affective, and cognitive aspects of family problem-solving interactions. Journal of Abnormal Psychology, 101, 495-504.

Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: W. W. Norton.

Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: Harper & Row.

Younger, A. J., & Daniels, T. M. (1992). Children's reasons for nominating their peers as withdrawn: Passive withdrawal versus active isolation. Developmental Psychology, 28, 955-960.

Younger, A., Schneider, B., Wadeson, R., Guirguis, M., & Bergeron, N. (2000). A behavior-based peer-nomination measure of social-withdrawal in children. Social Development, 9, 544-564.

Zupan, B. A., Hammen, C., & Jaenicke, C. (1987). The effects of current mood and prior depressive history on self-schematic processing in children. Journal of Experimental Child Psychology, 43, 149-158.

Kim B. Burgess

The George Washington University School of Medicine, Washington, D.C. and Pediatric Psychology Center, Rockville, MD

Alastair J. Younger

University of Ottawa 

Table 1
Self-Referent Rating Task: Positive
and Negative Descriptors

Positive      Negative
Descriptors   Descriptors

nice          bad
cheerful      boring
attractive    shy
smart         ugly
brave         immature
popular       lonely
leader        unhappy
good          ashamed
healthy       foolish
proud         nobody
happy         loser
funny         awful
exciting      sad
lively        embarrassed
helpful       angry
winner        moody
clever        afraid
friendly      terrible

Table 2
Correlations Among Measures of Self-Perception
and Behavioral-Emotional Problems for the Whole Sample

                    Negative     Internalizing   Externalizing
                      self-          score           score
                   perceptions

Positive             -.39 **        -.41 **          -.12
self-perceptions

Negative                            .46 **          .32 **
self-perceptions

Internalizing                                       .40 **
score

Note: For all correlations N = 132; * p < .05 and ** p <.01.

Table 3
Means and Standard Deviations for the Risk and
Control Groups on Self-Perception and Behavioral
Problem Measures

                  Number of        Number of
                positive words   negative words
                   endorsed         endorsed
Groups              M (SD)           M (SD)

Shy/withdrawn    12.85 (2.18)     3.10 (1.86)
Aggressive       15.07 (2.46)     2.02 (1.62)
Control          15.86 (2.12)     1.49 (1.19)

                Internalizing    Externalizing
                   problems         problems

Groups              M (SD)           M (SD)

Shy/withdrawn    56.48 (8.72)     48.63 (8.23)
Aggressive       51.07 (7.98)     56.59 (8.38)
Control          48.55 (8.86)     48.29 (9.33)

Table 4
Means and Standard Deviations for the
Socially Withdrawn and Normative Control
Groups on Each YSR Internalizing Subscale

               Shy/        Anxious/       Somatic
             Withdrawn     Depressed    complaints
Groups        M (SD)        M (SD)        M (SD)

Withdrawn   5.23 (2.54)   8.18 (4.76)   4.18 (3.01)
Control     2.96 (1.84)   4.45 (3.16)   2.88 (2.67)