Friends as reporters of children's and adolescents' depressive symptoms.

INTRODUCTION

Children and adolescents rarely initiate contact with mental health professionals. Instead, youth primarily receive assistance after being identified by others, usually parents and teachers (Clarizio, 1994). Thus, it is argued obtaining information from multiple sources is beneficial for understanding emotional functioning among youth (Achenbach, McConaughy, & Howell, 1987; Kazdin, 1994). Significant relations have been found between adults' reports of youths' emotional functioning and the youths' self-reports (Cantwell, Lewinsohn, Rohde, & Seely, 1997), but the magnitude of these relations is typically small to moderate (e.g., Achenbach et al., 1987). In addition, sibling reports (Epkins & Dedmon, 1999) and aggregate reports from classmates regarding depressive symptoms are related to children's self-reports; nonetheless, these relations are also small to moderate (e.g., Cole, Truglia, & Peek, 1997; Crowley, Worchel, & Ash, 1992). Accordingly, identifying additional possible reporters of youths' distress may be important for detecting at-risk youth.

Despite findings that those in close dyadic relationships (e.g., parents) and same-age classmates provide reports of depressive symptoms that are significantly related to youths' self-reports, close friends have not been considered as reporters of depression. This is unfortunate given friends' roles as companions and confidants (Rubin, Bukowski, & Parker, 1998). Accordingly, friends' reports of youths' depressive symptoms are the focus of this study. If relations between friend- and self-reports are especially strong, this research may have identified particularly valuable informants. Even if relations are comparable to those of other reporters, this study may nevertheless represent a step toward increasing our ability to identify at-risk youth.

Consider, as an example, the possibility that parents and friends are aware of different depressive symptoms. Parents are in a privileged position to observe symptoms such as eating or sleeping changes. However, youth disclose more to peers and less to adults with age (Buhrmester & Furman, 1987). As a result, among older children and adolescents, friends might be more knowledgeable than parents about symptoms of emotional distress (e.g., feeling sad or lonely). Friends are also in a unique position to observe symptoms related to social contexts, such as withdrawing or exhibiting a lack of pleasure when associating with peers. If parents and friends do have differential knowledge of depressive symptoms, future studies might reveal that some distressed youth who would not be identified using adult reports could be identified with friend reports. This study provides an initial test of this possibility by determining whether there are significant associations between friend- and self-reports of depressive symptoms.

This investigation also provides information regarding individual differences in youths' knowledge about friends. Although friendship researchers study self-disclosure in friendships (Buhrmester & Prager, 1995), little research has addressed how much youth actually know about friends. Guided by studies indicating individual differences in self-disclosure (e.g., Buhrmester & Furman, 1987; Parker & Asher, 1993), this study tests whether relations between friend- and self-reports of depressive symptoms are moderated by gender, grade, and friendship quality.

Parents, Teachers, Siblings, and Classmates as Reporters of Youths' Depressive Symptoms

Research regarding others' reports of youths' depressive symptoms typically involves adult informants, largely parents and teachers. As primary caregivers, parents are in a unique position to observe some depressive symptoms (e.g., sleeping habits), and parents' reports are related to their children's self-reports but the magnitude of these relations is small to moderate (e.g., Cantwell et al., 1997; Renouf & Kovacs, 1994; see Achenbach et al., 1987). Teachers' reports are also significantly related to youths' self-reports of depressive symptoms (e.g., Crowley & Worchel, 1993; Youngstrom, Loeber, & Stouthamer-Loeber, 2000), but agreement between teacher and student reports is generally lower than between parent and child reports (Puura et al., 1998; Williams, McGee, Anderson, & Silva, 1989). Illustratively, Achenbach et al.'s (1987) meta-analytic review evidenced an average correlation of .25 between parent reports and youth self-reports of youth adjustment and an average correlation of .20 between teacher reports and youth self-reports of youth adjustment. Moreover, even lower agreement could be expected for depressive symptoms given that these agreement statistics were based on reports of botth internalizing (e.g., depression) and externalizing (e.g., hyperactivity) problems, and greater agreement is found for externalizing than internalizing symptoms. Although a strength of adults as informants is that their mature cognitive capabilities and greater life experience should help them gauge the degree to which youth are suffering, adults have little access to some contexts relevant to the development or expression of depressive symptoms, namely those in which only youth are included.

In comparison to parents and teachers, siblings are likely to participate in certain contexts from which adults are generally excluded but in which depressive symptoms may be expressed. For example, siblings may participate in contexts such as neighborhood playgroups and riding the school bus. Siblings are rarely considered as reporters of emotional functioning, but one study did reveal significant relations between siblings' and children's reports of depressive symptoms (r = .22; Epkins & Dedmon, 1999). A challenge in studying siblings as reporters, however, is that they are a heterogeneous group. For instance, greater agreement may be found for siblings of the same sex who are similar in age and spend considerable time together than for siblings lacking these characteristics.

Classmates also have access to some contexts from which adults and siblings generally do not, such as the lunchroom and after-school activities. Many studies indicate that peers effectively report on classmates' social behavior (Rubin et al., 1998), and a smaller body of research has considered peers as reporters of emotional adjustment (e.g., Lefkowitz & Tesiny, 1980; Lefowitz, Tesiny, & Solodow, 1989). The most commonly used approach for utilizing peers as reporters of depression involves students nominating classmates for items describing depressive symptoms (e.g., "Who often looks sad?"). Aggregate scores based on classmates' nominations are related to self-reported symptoms with a magnitude comparable to those found using adult reports (e.g., r = .23 in Lefkowitz & Tesiny, 1980; see Achenbach et al., 1987).

A disadvantage of this approach is that the method does not take into account variation among classmates' relationships in that each peer's nomination is assumed to be as valid as every other. Youth seek out close friends for support and assistance (e.g., Boldero & Fallon, 1995), and peer reports may be more useful when provided by friends. Also, the effectiveness of aggregate peer-report measures has been demonstrated primarily with early elementary school children (e.g., Crowley et al., 1992; Lefkowitz & Tesiny, 1980, 1985; cf., Worchel, Hughes, Hall, Stanton, & Little, 1990). Obtaining reports from friends may be especially important for older youth as dyadic friendships are thought to assume greater importance with age (Parker & Gottman, 1989).

Friends as Reporters of Children's and Adolescents' Depressive Symptoms

This study is the first to test whether friends' reports of depressive symptoms are related to youths' self-reports. Little research addresses youths' knowledge of friends, but one study indicates that youth are relatively knowledgeable about friends' personality characteristics (Diaz & Berndt, 1982). Perhaps youth also are aware of friends' depressive symptoms. Further, this study examines whether the relation between friend- and self-reports depends on the type of symptom. Greater concordance might be expected for more easily observed behavioral depressive symptoms than internally experienced affective symptoms or somatic symptoms.

Gender

Gender may moderate relations between friend- and self-reports of depressive symptoms. Specifically, friend- and self-reports may be related more strongly for girls than boys. Girls self-disclose to friends (e.g., Parker & Asher, 1993) and seek …