Privileged but Pressured? A Study of Affluent Youth

Abstract

The purpose of this study was to build on preliminary findings of unusually high internalizing symptoms and substance use among suburban high school students. The sample consisted of 302 sixth- and seventh-grade students in an affluent, suburban community. Findings corroborated expectations regarding several domains of vulnerability, showing (1) high rates of clinically significant depressive symptoms among older girls, (2) significant links between various internalizing symptoms and substance use among both boys and girls, and (3) peers’ approval of substance use among older boys. In exploring potential causes of distress in this suburban sample, associations were found for achievement pressures (particularly excessive perfectionistic strivings), and isolation from parents (particularly low perceived closeness to mothers). Findings of this study are discussed in terms of widespread stereotypes about affluent families, as well as implications for future research and preventive interventions for a subgroup of youth typically viewed as being at “low risk.”

INTRODUCTION

Although developmentalists have increasingly emphasized the importance of contextual factors in child adjustment (Cicchetti & Toth, 1995; Sameroff, Seifer, & Bartko, 1997), there remains unevenness in the attention accorded to different subgroups. With growing acknowledgment of the limits of “one-model-fits-all” approaches, recent years have seen a surge in efforts focused on economically disadvantaged and minority youth (see Garcia Coll et al., 1996; Huston, 1994; Luthar, 1999). In contrast, there has been little explicit consideration of the unique life experiences of children at the opposite end of the socioeconomic spectrum: those in affluent, suburban families.

The relevance of context-sensitive research on suburban children is underscored by various suggestions of adjustment problems among the relatively affluent. Epidemiological research has established higher rates of depression in developed countries than in others (Buss, 2000), and historical data over the last 3 decades indicate that Americans are twice as rich now but no happier than they used to be. In the meanwhile, divorce rates have doubled, teen suicide has tripled, and depression rates have soared, especially among teens and young adults (Diener, 2000; Meyers, 2000). Mechanisms posited to explain such links have included those resting on individuals’ rapid “habituation” to new wealth and subsequent hankering for more; envy of people with the most perceived successes; and intense emotional isolation spawned by resolute pursuit of personal ambitions. As Csikszentmihalyi (1999, p. 823) has argued, “… to the extent that most of one’s psychic energy becomes invested in material goals, it is typical for sensitivity to other rewards to atrophy.”

There are some preliminary suggestions that such trends, documented among affluent adults, may generalize to children as well. For example, a study of almost 1,000 American teens revealed a low negative relation between parents’ socioeconomic status (SES) and teenagers’ self-reported happiness (Csikszentmihalyi & Schneider, 2000). Problems spanning multiple domains of adjustment were detected in a recent comparative study involving over 250 upper SES, suburban tenth graders, and their economically disadvantaged, inner-city counterparts. Suburban youth were found to report significantly higher levels of anxiety symptoms and cigarette, alcohol, marijuana, and hard drug use (Luthar & D’Avanzo, 1999). Comparisons with national normative samples yielded further signs of trouble. More than one in five suburban girls (22%) reported clinically significant depressive symptoms—rates three times as high as those in normative samples (7%). Rates of clinically significant anxiety among both boys and girls were somewhat higher than norms (22% and 26%, versus 17%, respectively). Frequency of substance use was also escalated, particularly alcohol use among girls (72% during the past year versus 61% in normative samples), and illicit drug use among boys (rates of 59% versus 38%, respectively).

Aside from indicating unusually high self-reported problems among suburban teens, this comparative study (Luthar & D’Avanzo, 1999) also illuminated two sets of risk factors for substance use, that seemed relatively specific to them. First, substance use was linked with self-reported maladjustment among affluent teens but not among inner-city students, suggesting that the former may often have used substances in attempts to “self-medicate” (see also Way, Stauber, Nakkula, & London, 1994). Second, the peer group in the suburban school seemed to endorse substance use among boys. Peer popularity was linked with high substance use among upper SES boys (but not among inner-city boys, or either group of girls), and this association remained robust even after controlling for various potential confounds in multivariate analyses.

The present study was implemented with a sample of middle school youth in an effort to build on these preliminary findings with suburban high school students. There were two major research objectives: The first was to replicate Luthar and D’Avanzo’s (1999) findings on (1) high distress and substance use, and (2) the potential role of distress as well as peer endorsement in relation to risk for students’ substance use. Replicatory evidence on these issues was seen as important from the perspective of ascertaining the need for, and potential directions of, any future preventive efforts that target this subgroup of youngsters. The second objective was to begin to explore contextually salient forces that might contribute to high maladjustment among affluent, suburban youth, as is elaborated in the section that follows.

Possible Causes of High Distress

Although developmental research that is focused on affluent children is still nascent—precluding coherent “theories” on contextually salient pathways in adjustment (see Garcia Coll et al., 1996)—work in other disciplines, such as sociology and clinical psychology, yield some useful leads. Supplementing these are insights derived from our own work in suburban communities over time. Using a strategy commonly employed in qualitative, ethnographic research (LeCompte & Preissle, 1993), we interviewed several “key informants” (members of the community likely to have a good understanding of processes within that community) about potential causes of students’ distress.¹ Respondents included school administrators, clinical service providers in the community, PTA representatives, and several parents and children. Insights gained from these various sources consistently pointed to two major factors as possible causes of distress, both of which were targeted for empirical scrutiny within this study.

The first of these potential causes of distress was excessive achievement pressures. In upwardly mobile suburban communities, there is often a ubiquitous emphasis on ensuring that children secure admission to stellar colleges. As a result, many youngsters feel highly driven to excel not only at academics but also at multiple extracurricular activities, with these pressures beginning as early as the middle school years (Luthar & D’Avanzo, 1999).

In examining the role of achievement pressures in the present study, the construct of achievement pressures was operationalized in terms of two dimensions, the first of which encompassed parental emphasis on achievements. As Alice Miller (1997) noted in her clinical treatise, The Drama of the Gifted Child, when children come to believe that their parents value them more for what they can do than for who they are, they tend to rely on their accomplishments for their sense of self-worth. This, in turn, places them at high risk for maladjustment, because of intense preoccupation with real, imagined, or anticipated failures (see also Bing, 1999). Similarly, empirical studies with adults have shown that those who place strong importance on extrinsic goals (e.g., wealth and image) relative to intrinsic goals (such as affiliation and personal growth) tend to be at heightened risk for depression and anxiety (Kasser & Ryan, 1993, 1996).

The second dimension used to operationalize achievement pressures was children’s own maladaptive perfectionistic strivings. Previous research on perfectionism has indicated two distinct subtypes of the construct: an adaptive subtype that represents the tendency to organize one’s work and set appropriately high standards for the self, and a maladaptive subtype connoting excessive investment in accomplishments and need to avoid failure. Furthermore, maladaptive (but not adaptive) perfectionism has been shown to be highly correlated with indices of distress, such as depression and anxiety (Flett, Hewitt, & Dyck, 1989; Hewitt, Flett, & Turnbull, 1992).

Aside from achievement pressures, a second broad factor potentially salient in suburban students’ adjustment disturbances is disconnection from adults—both literal and emotional (Luthar & D’Avanzo, 1999). Sociological analyses of contemporary American family life indicate that among many upper-middle-class families, junior high school students often have little or no after-school supervision by adults (Hochschild, 1997). Most often, this phenomenon reflects not a lack of child care, but rather, parents’ beliefs that it promotes children’s self-sufficiency.

In a similar vein, it has been suggested that suburban children’s needs for emotional closeness with parents can suffer as the demands of professional parents’ careers erode relaxed “family time,” and youngsters are shuttled between various after-school activities (Hurley & Lustbader, 1997; Luthar & D’Avanzo, 1999). Consistent with this speculation, national survey data (U.S. Department of Health and Human Services, 1999) show that among 12- to 17-year-old youth, closeness to parents tends to be inversely linked with household income. Feelings of high closeness to resident biological mothers, for example, were reported by approximately 75% of youngsters whose annual family incomes were below $15,000, but only by 65% of youth whose family incomes were above $75,000. Comparable statistics for closeness to resident biological fathers were 66% and 54%, respectively.

It is also possible that closeness to parents, particularly mothers, can buffer many suburban children from other negative forces. To illustrate, it may not be very deleterious for 12-year-olds to be alone at home after school, if they generally experienced their parents as being highly aware of and involved with their everyday activities. The high protective potential of closeness to mothers, in particular, is suggested by common patterns of child-care responsibilities. In affluent communities, mothers tend to be more actively involved in home-based activities than are fathers, because men’s professional careers often necessitate their frequent absence from home (Mederer, 2000).

Prior evidence based on other sociodemographic samples supports the consideration of each of these socializing influences as possible causes of apparently high levels of distress in suburban youth. Achievement pressures, for example, have been linked with high symptom levels and substance use (Ablard & Parker, 1997; Frost, Lahart, & Rosenblate, 1991; Steinhausen & Metzke, 1998), as has inadequate parental monitoring (Kandel & Davies, 1996; Kazdin, 1995; Zucker, Fitzgerald, & Moses, 1995), and, in particular, lack of supervision after school (Cooper, Valentine, Nye, & Lindsay, 1999; Mulhall, Stone, & Stone, 1996; Richardson, Radziszewska, Dent, & Flay, 1993). Similarly, the protective role of feelings of closeness to parents has been amply documented (Hawkins, Catalano, & Miller, 1992; Masten, 2001); more specifically, closeness to mothers has been shown to be a major determinant of well-being among children and adolescents (Frank, Pirsch, & Wright, 1990; Sheeber, Hops, Alpert, Davis, & Andrews, 1997).

Additional Parameters of This Study

Apart from exploring these various socializing influences (achievement pressures and isolation from parents) in relation to suburban children’s self-reported maladjustment, implications for their academic grades were also considered in the present study. Although ubiquitous pressures to achieve may in fact exacerbate students’ distress levels, they might, equally, bolster motivation and the levels of success actually attained at school. Accordingly, students’ grades were examined as outcomes in addition to self-reported adjustment problems.

All analyses were conducted separately by gender, given substantial evidence of gender differences in the effects of risk factors such as those examined here (Blatt & Zuroff, 1992; Frome & Eccles, 1998; Gjerde, Block, & Block, 1991; Luthar & D’Avanzo, 1999). Finally, in the interest of detecting links specific to the risks under consideration, analyses entailed controls for children’s perceived difficulties with peers. This was intended to correct for any artifactual inflation of associations deriving from the measurement of major predictors and outcomes via self-reports. Subjectively perceived peer victimization was selected as a control variable (rather than any other self-reported problem) because of (1) high salience during the developmental period of preadolescence (Crick & Grotpeter, 1996, 1998), and (2) its potential to partial out any variance arising from negative views of relationships in general, as opposed to those specific to relationships with parents (salient among the predictor variables).

In sum, the central objective of this study of sixth-and seventh-grade students was to build on prior evidence that has suggested areas of surprisingly high maladjustment among youth in upper-class suburban families. Objectives were to ascertain the general-izability of findings previously obtained with high school students, on (1) the incidence of clinically significant depressive and anxiety symptoms relative to normative rates, as well as (2) links between substance use and indicators of both symptomatology and school-based competence. Yet another objective was to explore factors that might explain high maladjustment among suburban youth, including achievement pressures (excessive perfectionism and perceived parental emphasis on achievements), and dimensions of isolation from adults (lack of after-school supervision and low closeness to mothers and fathers).

METHOD

Sample

Participants in this study consisted of sixth- and seventh-grade students who were attending a middle school in an affluent community in the Northeast. During the year 2000, the median annual family income in this region was reported to be almost $102,000; the highest national median income ever recorded by the U.S. Census is $40,816 (U.S. Census Bureau, 1999). Of the total sample of 302 students, 168 were sixth graders, and 134 were in the seventh grade, with mean ages of 11.8 and 12.8 years, respectively. Most students (92%) were of European American background; of the rest, 1.5% were African American, 1.5% were Hispanic American, 3% were Asian American, and 2% were of other ethnic backgrounds.

This study was implemented as part of a school-based initiative that targeted positive youth development, and inclusion of students in the sample was based on passive consent procedures. Given increasing anecdotal and media-based evidence of various problems in communities such as theirs, school administrators and parent representatives in this particular community had sought a rigorous assessment of the nature and extent of difficulties among their middle school students. Following the development of a collaborative relationship with the first author (which ensued from a series of talks for the community), the survey was initiated. School administrators sent letters to parents of all sixth and seventh graders by U.S. mail that described the project, indicated that survey results would be presented only in aggregate form (with no information on individual children), and requested notification if they preferred that their children not participate. A second notice was mailed a few days before data collection, once again offering parents the option to refuse consent. On each of the 2 days of data collection, all students were also told that their participation was entirely voluntary. On completion of data collection, questionnaires were scored with only participant numbers as identifiers.

Complete data were obtained for 95% of the sixth and seventh graders who were attending the school sampled. Of the 16 students who were not included, parents of 8 did not wish for them to participate, 2 children themselves did not want to participate, 4 were excluded because of incomplete data, and 2 were absent on both days of data collection.

Measures: Subjective Reports of Maladjustment

Depressive symptoms

The Children’s Depression Inventory (CDI; Kovacs, 1992) is a widely used 27-item, three-choice scale designed for school-age children and adolescents. This measure has acceptable internal consistency as well as criterion and concurrent validity (Kovacs, 1992). Internal consistent coefficients in this sample were .90 and .89 for girls and boys, respectively.

Anxiety

The Revised Children’s Manifest Anxiety Scale (R-CMAS; Reynolds & Richmond, 1985) is a dichotomous-choice, 37-item, self-report measure. The instrument provides scores on three dimensions of anxiety (social anxiety, physiological anxiety, and worry) as well as a total anxiety score based on all of these collectively. Acceptable reliability and validity coefficients have been reported for R-CMAS scores (Luthar & D’Avanzo, 1999; Reynolds & Richmond, 1985). In this sample, Cronbach’s α coefficients were .86 for total anxiety among girls and .88 among boys; coefficients for the subscales were .65, .74, and .82 for girls, and .63, .75, and .79 among boys.

Delinquent behavior

Participants completed the Self-Report Delinquency Checklist (SRD; Elliot, Dunford, & Huizinga, 1987), an instrument in which students are asked about the occurrence of delinquent acts at home, at school, and in the community. The measure includes 37 items rated on a 4-point scale, and the total score on the SRD reflects the frequency of delinquent behavior. The SRD has been shown to be a valid and reliable instrument (Huizinga & Elliot, 1986).

Following Luthar and D’Avanzo’s (1999) scoring procedures, six items on the SRD that pertained to substance use (e.g., “used alcohol”) were removed before computing total delinquency scores, to avoid redundancy in examining statistical links with substance use. Alpha coefficients for the resultant SRD scores were .92 and .94 for girls and boys, respectively.

Substance use

The frequency of drug use grid that was used in the Monitoring the Future Study Survey (Johnston, O’Malley, & Bachman, 1984) was administered to students in the present sample. This instrument queries about frequency of use of several substances—nicotine, alcohol, marijuana, inhalants, crack, cocaine, and LSD—over the preceding year with ratings obtained on a 7-point scale anchored by “never” to “40+ times.” The reliability and validity of this type of self-report have been amply documented (Johnston, Bachman, & O’Malley, 1989; Wallace & Bachman, 1991). Following Luthar and D’Avanzo’s (1999) approach, for major statistical analyses in this study, a composite substance use variable was created by adding scores for nicotine, alcohol, and marijuana; this variable was logarithmically transformed to reduce skewness and kurtosis. Alpha coefficients for this variable were .66 and .83 for girls and boys, respectively.

Pressures to Achieve

Parent values

To ascertain the degree to which students perceived parents as emphasizing their achievements, a Parental Values measure was developed that consisted of a list of 10 issues potentially important to parents. Students were asked to rank order the top five of the following items based on what they believed their own parents valued the most: “that you … are respectful to others, attend a good college, excel academically, make a lot of money in the future, are a leader in school, shine in extracurricular activities, are honest, always try to help others in need, are kind to others, and are generally happy with yourself and your life.” In an initial examination of the validity of this instrument, latent class analyses based on ranking revealed three meaningful clusters, one of which consisted of students who saw their parents as valuing their achievements highly, and the other two consisting of children whose parents valued their personal character and well-being more than achievements (DeCarlo & Luthar, 2000). For future analyses based on this measure, DeCarlo and Luthar (2000) indicated the utility of either latent class analyses or simple tallies of achievement items endorsed in the top five items; the latter scoring approach was used in the present study.

Perfectionism

The Multidimensional Perfectionism Scale (MPS; Frost, Marten, Lahart, & Rosenblate, 1990) is an instrument with 35 questions rated on a 5-point Likert scale, and yields six subscale scores (illustrative items are listed in parentheses): concern over mistakes (“The fewer mistakes I make, the more people will like me”), doubts about actions (“If someone does a task at work/school better than I, then I feel like I failed the whole task”), parental expectations (“My parents expect excellence from me”), parental criticism (“my parents never try to understand my mistakes”), personal standards (“I set higher goals than most people”), and organization (“Neatness is very important to me”). In this sample, sub-scale αs ranged from .62 to .91 (median = .81) for girls, and from .62 to .94 (median = .74) for boys.

Isolation from Adults

After-school supervision

Students indicated the type of supervision that they typically received after school based on six possible choices: (1) supervised by parent, (2) supervised by other adult (nonrelative), (3) home with sibling but no adult present, (4) organized social activity with no adult present, (5) with a friend but no adult present, and (6) home alone.

Parent–child closeness

The quality of parent–child relationships was assessed by two questionnaires. The Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987) consists of 50 items (25 each pertaining to mothers and to fathers), and assesses the degree of trust, communication, and alienation in relationships with parents (e.g., “My mother understands me,” “My mother can tell when I’m upset about something,” “I feel angry with my mother”). Items are rated on a 4-point Likert scale. In this sample, Cronbach’s α coefficients for the six subscales ranged between .76 and .91 for girls, and .74 and .93 for boys.

Parental involvement was assessed by a series of 16 questions derived from questionnaires used in previous studies (Dornbusch, 1985; Patterson & Stouthamer-Loeber, 1984). Ten questions asked about interactions between children and their parents (e.g., how often children talked with and did things with their parents). Another six questions asked about parents’ awareness of their children’s everyday activities (e.g., how much parents knew about how and with whom their children spent their time). Questions were rated on a 3-point scale, and were asked separately for mothers and fathers; α coefficients in this sample were .76 and .85 respectively, for girls, and .74 and .83 for boys.

Academic Iindices

School grades and absences

Data on these variables were obtained from student records. An end-of-year cumulative grade-point average was computed based on students’ grades across the first three quarters of the school year (the 4th quarter was still in session at the time of data collection). This cumulative score was based on averaged grades in social studies, science, and math, as well as language arts (for sixth graders) or English (for seventh graders); letter grades were coded such that an an A+ was assigned a score of 13 and an F was assigned a score of 1.

Teacher ratings

The Teacher–Child Rating Scale (T-CRS; Hightower et al., 1986) was given to English teachers of all students. A 36-item scale, the T-CRS assesses behaviors within two domains with three scores within each: Problems (acting out, learning problems, and shy-anxious) and competence (frustration tolerance, assertive social skills, and task orientation). All scores are standardized to correct for individual teachers’ rating patterns. Acceptable psychometric properties have been reported for this measure (High-tower et al., 1986; Luthar, 1995). Alpha coefficients for the Problem Scale were .92 for both girls and boys, and for the Adjustment Scale were .94 and .95, respectively.

Peer Indices

Peer victimization

The Social Experience Questionnaire (Crick & Grotpeter, 1996) contains 15 items on experiences with peers, each rated on a 5-point scale. The instrument encompasses two dimensions of victimization, overt victimization (e.g., “How often do you get pushed around or shoved by another kid at school?”), and relational victimization (e.g., “How often do other kids leave you out on purpose when it is time to play or do an activity?”). A third subscale assesses receipt of prosocial acts (e.g., “How often does another kid do something that makes you feel happy?”). The two victimization scores were of direct relevance to the present analyses (given the intention to control for biases due to self-reports on negative indices). These two scores were summed to derive a composite victimization score, which had α coefficients of .85 and .86 for girls and boys, respectively.

Peer status

Following Luthar and D’Avanzo (1999), Coie, Dodge, and Coppotelli’s (1982) sociometric procedure was used to ascertain peer acceptance and rejection. This procedure entails asking children to list up to three students whom they like the most within their class, and up to three they like the least. Students were allowed to nominate classmates of either gender. The total number of positive and negative nominations that each student received was standardized by class (to account for varying classroom sizes) for inclusion in statistical analyses. The reliability and validity of this sociometric method have been amply documented (Coie & Dodge, 1983; Coie, Dodge, & Kupersmidt, 1990; Coie, Lochman, Terry, & Hyman, 1992).

Procedure

Data for each student were collected during two 45-min class periods on two separate days, and testing of the children was done in groups of 15 to 20. To guard against biases due to variability in reading proficiencies, a member of the research team read each questionnaire aloud, and students marked their responses accordingly. Questionnaires were administered in the same order to all groups, with relatively structured, nonthreatening measures administered at the beginning and end of each session. On completion of data collection, a gift of money to support a pizza party was given to all participating classes, and teachers were given $1 for each student they rated.

RESULTS

Data Reduction

In view of prior research that indicated two dimensions of perfectionism, a principal components factor analysis with varimax rotation was performed on MPS subscale scores. A two-factor solution was indicated by eigenvalues greater than 1. This solution accounted for 99% of the variance. The first factor reflected a maladaptive perfectionism, with high loadings on concern over mistakes, doubts about actions, parental expectations, and parental criticism. The second dimension depicted an adaptive perfectionism, with high loadings on personal standards and organization. Similar dimensions were found in subgroups separated by gender and by grade level in this sample, as well as in prior research that involved over 800 academically gifted sixth graders (Parker & Stumpf, 1995). Two dimensions were therefore derived: maladaptive perfectionism (αs = .87 and .85 for girls and boys, respectively) and adaptive perfectionism (αs = .83 and .88, respectively); the former was relevant to analyses in the present study.

To reduce the number of variables pertaining to relationships with mothers and fathers, factor analyses with varimax rotation were conducted on attachment and involvement subscale scores for each parent. Results indicated one main factor, which accounted for 64% of the variance for mothers and 60% for fathers. Accordingly, two composite scores, mother closeness and father closeness, were derived, based on the following combination of standardized scores: (trust + communication + involvement − alienation). For mother closeness, α coefficients were .90 for both girls and boys, and for father closeness, they were .90 for girls and .87 for boys. Finally, given the consistently high correlations between scores on the CDI and the R-CMAS in other samples (Luthar, 1991, 1995; Luthar & D’Avanzo, 1999) as well as this one (r = .70), standardized depression and total anxiety scores were summed to form an overall emotional distress variable; α coefficients were .88 for both girls and boys.²

Descriptive Statistics

Means and standard deviations on all variables are presented in Table 1, separately by grade and gender. A two-way multivariate analysis of variance (Gender × Grade level) indicated significant main effects for both, as well as a significant interaction effect. Follow-up univariate analyses indicated that girls fared better than did boys in most spheres, as did sixth graders relative to seventh-grade students.

Table 1.

Descriptive Data on All Variables Presented by Gender and Grade Level

	Sixth Grade				Seventh Grade				Fgender	Fgrade level
	Girls (n = 74)		Boys (n = 94)		Girls (n =67)		Boys (n = 67)
	M	SD	M	SD	M	SD	M	SD
Depression	9.01	5.90	7.10	5.54	10.63	6.38	7.02	5.47	.92	2.74
Anxiety	8.50	6.90	7.52	6.81	9.82	8.40	9.13	8.29	15.50***	1.19
Distressa	−.01	.22	−.34	.19	.52	.23	−.11	.23	4.92*	3.05
Delinquency	1.15	.04	1.24	.03	1.15	.04	1.55	.04	32.34***	17.72***
Substance use	.08	.00	.02	.00	.02	.01	.05	.01	16.90***	25.50***
Grades	10.07	.27	9.73	.23	9.25	.28	7.33	.28	18.48***	37.58***
Teacher-rated competence	.61	.29	.70	.30	−.18	.25	−.95	.30	18.17***	1.38
Teacher-rated problems	−.46	.25	.33	.22	−.81	.26	.65	.26	20.26***	.00
Days absent from school	6.90	.65	5.66	.56	6.91	.67	5.51	.68	4.24*	.01
Maladaptive perfectionism	−.38	.34	−.18	.30	.35	.36	.31	.36	.05	3.22
Parent achievement emphasis	3.81	.43	3.70	.37	3.54	.44	4.72	.45	1.63	.79
Mother closeness	.40	.32	.32	.28	.20	.33	−1.05	.34	4.28*	6.03*
Father closeness	.02	.31	.76	.26	−.68	.32	−.29	.32	3.43	8.31**
Liked least	−.02	.12	.04	.10	−.18	.12	.15	.13	2.90	.04
Liked most	.03	.12	−.10	.10	.23	.12	−.23	.12	6.67**	.11
Type of after-school supervision (% of students)
Adult family member	50.0		37.2		46.2		39.7
Adult	23.4		12.0		26.2		11.1
Alone	26.6		25.0		27.7		49.2

Note: Gender × Grade interaction terms were significant in two cases: delinquency, F = 16.02***, and grades, F = 9.11**.

^aComposite integrating standardized scores on both depression and anxiety.

^*p < .05;

^**p < .01;

^***p < .001.

Also displayed in Table 1 are proportions of children in different groups based on after-school supervision by adults. Between one third and half of sixth and seventh graders were usually supervised by their parents. However, no adult supervision was reported by one quarter each of the sixth-grade girls, sixth-grade boys, and seventh-grade girls, and by almost half of the seventh-grade boys in this sample, χ² = 6.27, p < .01.

Given the dearth of evidence on the implications of different types of after-school supervision among suburban children, variations in adjustment were explored across the three groups depicted in Table 1. Multivariate analyses of variance indicated significant effects across supervision groups, Wilks’s λ, 2.21, p < .001 (but nonsignificant effects for interactions of group with gender and with grade level). Post hoc comparisons showed that across all adjustment indicators, differences lay between means of the first two groups on the one hand and the third on the other. Based on these findings, a dichotomous distinction was derived for subsequent statistical analyses—supervision by any adult versus no adult supervision.

Simple correlations of all variables are presented in Table 2. Values for sixth-grade boys and girls are presented above the diagonal, and for seventh-graders, below the diagonal.

Table 2.

Intercorrelations among Variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Distress		.54**	.33**	.00	.55**	.15	−.20	−.12	−.32**	.03	−.25*	.27*	.42**	−.03	.15
		.41**	.33**	.00	.43**	.01	−.36**	−.23*	−.18	.02	−.18	.17	.36**	.16	−.07
2. Delinquency	.60**		.37**	−.09	.47**	.18	−.14	−.13	−.26*	.14	−.11	.14	−.43**	.09	.08
	.40**		.59**	−.03	.18	.10	−.09	−.19	−.47**	.13	−.34**	.37**	−.12	.06	.08
3. Substance use	.29*	.59**		−.16	.20	.19	−.23	−.18	−.32**	.27*	−.11	.03	.09	.05	−.18
	.52**	.78**		−.12	.06	.03	−.11	−.20*	−.26*	.24*	−.28**	.29**	.09	.26**	.02
4. Parent achievement emphasis	.41**	.15	−.12		.20	.08	−.31**	−.25*	−.10	−.06	−.13	.16	.09	.15	−.21
	.04	−.10	.03		.36**	.21*	−.11	.02	−.07	−.09	−.12	.17	.05	.24*	−.11
5. Maladaptive perfectionism	.56**	.32**	.13	.47**		.21	−.20	−.15	−.21	.00	−.09	.13	.50**	.11	.08
	.45**	.15	.33**	.32**		.06	−.21*	−.05	−.03	−.10	−.00	.06	.19	.17	−.21*
6. No after-school supervision	.39**	.25*	.29*	.22	.11		−.26*	−.33**	.00	−.16	.07	−.13	.20	.20	−.29*
	.07	.13	.15	.02	.17		−.22*	−.28**	−.15	.09	−.20	.20	−.03	.10	.01
7. Mother closeness	−.57**	−.29*	−.32**	−.17	−.54**	−.22		.49**	.14	.13	.18	−.20	−.07	−.04	.07
	−.22	−.49**	−.49**	−.07	−.26*	−.47**		.57**	.06	−.05	.09	−.11	−.00	−.01	.12
8. Father closeness	−.37**	−.26*	−.15	−.20	−.20	−.31*	.52**		.19	−.10	.05	−.01	−.05	.16	.06
	−.09	−.31*	−.25*	−.03	−.15	−.34**	.57**		.06	−.11	.20*	−.22	−.07	−.13	.04
9. Grades	−.21	−.25*	−.19	−.19	−.23	−.13	.13	.37**		−.28*	.69**	−.65**	−.28*	−.25*	.18
	−.38**	−.17	−.24	.00	.01	−.05	.22	.16		−.20*	.73**	−.72**	.15	−.26**	−.02
10. Absences	.12	.04	.16	.27*	.24*	.15	−.15	−.05	.03		−.14	.12	−.18	.00	.02
	.36**	.04	.24	.24	.21	−.09	−.06	.22	−.31**		−.25*	.36**	.02	.17	.03
11. Teacher-rated competence	−.28*	−.35**	−.04	−.29*	−.17	−.05	.14	.31**	.75**	.10		−.80**	−.12	−.27**	.37**
	−.30*	−.07	−.12	.09	−.10	.00	.23	.19	.73**	−.21		−.83**	−.14	−.41**	.26*
12. Teacher-rated problems	.28*	.32**	.06	.18	.10	.10	−.02	−.30**	−.66**	.02	−.82**		.12	.16	−.21
	.35**	.27*	.27*	−.10	−.00	−.01	−.24	−.14	−.75**	.24*	−.81**		−.07	−.38**	−.03
13. Victimization by peers	.35**	.17	.34**	.10	.25*	.07	−.23	−.04	−.14	−.07	−.05	.15		.16	.05
	.47**	.11	.12	−.17	.17	−.20	.18	.21	−.15	.08	.09	.26*		.32**	−.21*
14. Liked least	−.06	.07	.15	.02	−.12	.10	.13	−.04	−.22	−.16	−.23	.22	.16		−.32**
	.06	.17	.12	−.01	−.01	.02	−.19	−.09	−.36**	.15	−.31**	.41**	.13		−.25*
15. Liked most	−.13	−.12	.03	.00	−.03	−.10	.18	.18	.25*	.17	.25*	−.23	−.14	−.24
	.05	.26*	.25*	.02	.17	−.03	.02	−.16	.16	−.16	.08	−.08	−.16	−.23

Note: Correlations for sixth graders are listed in the top half of the diagonal and correlations for seventh graders are listed in the bottom half of the diagonal, with girls listed in the top row and boys underneath.

^*p < .05;

^**p < .01.

Replication of Prior Findings with Suburban High School Students

Incidence of clinically significant symptoms

Following Luthar and D’Avanzo (1999), the proportion of students with clinically significant symptoms was examined on two indices for which normative data are available: depression and anxiety. As shown in Table 3, elevations in depression rates were apparent only among seventh-grade girls. On the CDI, 14% of these girls reported symptoms above the clinical cutoff (raw score = 19), as compared with 7% of girls in normative samples. Seventh-grade girls also stood out from the other subgroups on anxiety symptoms, with 1 in 5 reporting clinically significant anxiety symptoms. Rates of clinically significant anxiety were substantially below norms among sixth-grade girls as well as sixth- and seventh-grade boys.

Table 3.

Self-Reported Problems among Students in This Sample: Incidence of Clinically Significant Depressive and Anxiety Symptoms (as Compared with National Norms)

	Norms: 7–12-year-old Girls	Sixth-Grade Suburban Girls	Norms: 7–12-year-old Boys	Sixth-Grade Suburban Boys	Norms: 13–17-year-old Girls	Seventh-Grade Suburban Girls	Norms: 13–17-year-old Boys	Seventh-Grade Suburban Boys
Depressiona
Clinical cutoff: Above average
Raw score (T score 65)	20		22		19		24
Percentage of sample above cutoff	8	7	7	4	7	14	7	10
Clinical cutoff: Much above average
Raw score	23		25		22		28
Percentage of sample above cutoff	5	4	5	3	3	10	2	7
	Norms: 12-year-old Girls	12-year-old Suburban Girls	Norms: 12-year-old Boys	12-year-old Suburban Boys	Norms: 13-year-old Girls	13-year-old Suburban Girls	Norms: 13-year-old Boys	13-year-old Suburban Boys
Total anxietyb
Raw score	20		17		17		15
Percentage sample above cutoff	16	4	17	4	17	21	16	8

^aNormative data on depressive symptoms are based on values on the Children’s Depression Inventory (Kovacs, 1992).

^bNorms for total anxiety symptoms correspond to values for the Revised-Children’s Manifest Anxiety Scale (Reynolds & Richmond, 1985).

Whereas Luthar and D’Avanzo (1999) also compared substance use rates of their sample with national norms, this was precluded here because sixth- and seventh-grade students were not included in the national Monitoring the Future Study. Descriptive data on use by gender and grade level are, however, displayed in Table 4. As shown in this table, across various indicators of substance use and among both boys and girls, frequency rates were almost three times as high among seventh graders as opposed to sixth graders.

Table 4.

Self-Reported Substance Use Frequencies among Students in This Sample

Grade Level	At Least Once in Preceding Year				Once or More per Month on Average
	Sixth		Seventh		Sixth		Seventh
	Girls	Boys	Girls	Boys	Girls	Boys	Girls	Boys
Cigarettes	8	7	24	20	1	1	3	11
Alcohol	11	15	35	35	—	.2	4	18
Drinking to intoxication	1	6	9	28	—	—	—	7
Marijuana	—	—	6	7	—	—	—	7

Note: All values are percentages.

Factors associated with substance use

Following the approach used by Luthar and D’Avanzo (1999), hierarchical multiple regression analyses (Cohen & Cohen, 1975) were conducted to examine symptom indices and indicators of school-based behavioral competence as predictors of substance use. In each case, age was controlled for at the outset. As shown in Table 5, results showed that internalizing symptoms as a block were significantly related to substance use, with unique associations for physiological anxiety, social anxiety, depression, and delinquency.

Table 5.

Hierarchical Regression Analysis Predicting Substance Use from Self-Reported Symptom Indices

	Girls			Boys
	Step	β	R2Δ	Step	β	R2Δ
Age	1	.17	.03*	1	.26	.07**
Internalizing (block)			.14***			.23***
Physiological anxiety	2	.18	.03*	2	.28	.08***
Social anxiety	3	.22	.04*	3	.26	.05**
Worry	4	.01	.00	4	−.08	.00
Depression	5	.43	.06**	5	.43	.10***
Externalizing
Delinquency	6	.51	.15***	6	.65	.29***
Total			.32			.59

^*p < .05;

^**p < .01;

^***p < .001.

Analyses involving school-based behavioral indicators (shown in Table 6) indicated significant effects for the block of peer indicators among boys and for academic indices among both boys and girls. Peer acceptance and peer rejection each had unique positive links with substance use among boys. Negative associations were found for grades among both boys and girls, and positive links were found for boys’ teacher-rated problems and girls’ absenteeism.³

Table 6.

Hierarchical Regression Analysis Predicting Substance Use from Behavioral Social Competence Indices

	Girls			Boys
	Step	β	R2Δ	Step	β	R2Δ
Age	1	.17	.03	1	.29	.08***
Peers (block)			.01			.06**
Liked least	2	.08	.01	2	.20	.04*
Liked most	3	.00	.00	3	.15	.02*
Academics (block)	4		.12***	4		.10***
Teacher-rated problem behaviors	5	−.01	.00	5	.22	.04**
Teacher-rated positive behaviors	6	−.10	.00	6	−.02	.00
Grades	7	−.43	.09***	7	−.30	.04**
Days absent	8	.17	.03*	8	.13	.02
Total			.16			.24

^*p < .05;

^**p < .01;

^***p <.001.

Achievement Pressures and Isolation from Adults in Relation to Adjustment Indicators

Hierarchical regressions were performed to examine hypothesized links involving achievement pressures and isolation from adults. Again, in all analyses, age was entered at the outset. This was followed by feelings of peer victimization, entered to partial out any inflation of subsequent associations due to shared method variance.

Because adult socializing influences were of primary interest in the investigation, children’s reports on parent indices were given precedence in order of entry (i.e., over their own feelings of maladaptive perfectionism; Cohen & Cohen, 1975). Among the parent variables, closeness to mothers and closeness to fathers were entered first, given their well-established ramifications for child adjustment. These were followed by parental achievement values and after-school supervision. Within each of these two blocks, order of entry was allowed to vary depending on reducing tolerance (Tabachnick & Fidell, 1989). Finally, given suggestions of the significance of emotional closeness to mothers in particular, the degree to which this construct might moderate effects of other risk indices was explored via a set of three interaction terms: Mother Closeness × Parental Values, After-School Supervision, and Maladaptive Perfectionism.

Results of these analyses, presented in Tables 7 and 8 for girls and boys, respectively, showed significant associations for all predictor variables, with some showing effects across multiple outcomes. Closeness to mothers was associated with distress, delinquency, and substance use among both boys and girls, and additionally, to grades among boys. A single significant effect was found for closeness to fathers in relation to girls’ academic grades. After having considered variance due to these indices, the parental achievement values had a modest positive association with girls’ distress and was negatively linked with their substance use. After-school supervision had significant links with girls’ distress, delinquency, and substance use. Finally, maladaptive perfectionism was linked with distress and delinquency among both boys and girls.

Table 7.

Results of Hierarchical Regression Analyses for the Prediction of Girls’ Internal Distress, Externalizing Symptoms, and Academic Achievement

	Step	Internal Distress		Delinquency		Substance Use		Grades
		β	R2Δ	β	R2Δ	β	R2Δ	β	R2Δ
Predictors
Age	1	.06	.00	−.01	.00	.19	.04*	−.09	.01
Victimization by peers	2	.38	.14***	.26	.06**	.16	.03	−.13	.02
Mother closeness	3	−.41	.16***	−.19	.03*	−.29	.08**	.12	.02
Father closeness	4	−.07	.00	−.07	.00	.00	.00	.39	.11***
Parent achievement emphasis	5	.14	.02+	−.05	.02	−.23	.05*	−.05	.00
No after-school supervision	6	.15	.03*	.28	.07**	.22	.04*	.07	.01
Maladaptive perfectionism	7	.43	.11***	.28	.05*	.04	.00	−.08	.00
Interaction terms (block)a			.03		.09**		.03		.01
Mother Closeness ×
Parent Achievement Emphasis	8			−.14	.01
No After-School Supervision	9			−.33	.07***
Maladaptive Perfectionism	10			.09	.01
Total			.49		.31		.26		.17

^aAs recommended by Aiken and West (1991), interaction terms involve centered variables. Individual interaction terms within blocks were examined only if the block as a whole yielded a significant increase in R² change.

^*p < .05;

^**p < .01;

^***p < .001;

⁺p < .07.

Table 8.

Results of Hierarchical Regression Analyses for the Prediction of Boys’ Internal Distress, Externalizing Symptoms, and Academic Achievement

	Step	Internal Distress		Delinquency		Substance Use		Grades
		β	R2Δ	β	R2Δ	β	R2Δ	β	R2Δ
Predictors
Age	1	−.01	.00	.34	.12***	.28	.08**	−.23	.05**
Victimization by peers	2	.42	.17***	.13	.02	.18	.03*	−.07	.01
Mother closeness	3	−.32	.10***	−.38	.14***	−.36	.12***	.18	.03*
Father closeness	4	.02	.00	−.14	.01	−.14	.01	.01	.00
Parent achievement emphasis	5	−.01	.00	−.09	.01	−.06	.00	−.07	.01
No after-school supervision	6	.00	.00	−.03	.00	.01	.00	−.11	.01
Maladaptive perfectionism	7	.35	.10***	.18	.03*	.14	.02	.12	.01
Interaction terms (block)a			.03		.04*		.03		.03
Mother Closeness ×
Parent Achievement Emphasis	8			.28	.02*
No After-School Supervision	9			−.13	.02
Maladaptive Perfectionism	10			.04	.00
Total			.40		.36		.30		.14

^aAs recommended by Aiken and West (1991), interaction terms involve centered variables. Individual interaction terms within blocks were examined only if the block as a whole yielded a significant increase in R² change.

^*p < .05;

^**p < .01;

^***p < .001.

Interaction terms in the regression analyses were examined individually only if the block as a whole yielded a statistically significant increase in R² change (R²Δ). Results indicated two significant effects, both in relation to self-reported delinquency. In each case, low closeness to mother seemed to serve “vulnerability and reactive” effects (Luthar, Cicchetti, & Becker, 2000) in relation to the risks experienced. Thus, delinquency levels were relatively high among girls who lacked adult after-school supervision, and among boys who viewed their parents as emphasizing their achievements. In both instances, however, delinquency levels spiked sharply when the risk condition coexisted with low closeness to mothers.

Supplementary Analyses

To ascertain the degree to which order of entry of main effect terms might have biased findings, regression analyses were rerun with parental achievement values and supervision entered before mother closeness and father closeness. Results were similar, except that parental values had stronger links with girls’ distress, β = .23, R²Δ = .05, p < .01, whereas links with their substance use lost statistical significance, β = −.15, R²Δ = .02, ns. Similarly, links involving after-school supervision were stronger in relation to girls’ distress, β = .22, R²Δ = .05, p < .01; delinquency, β = .32, R²Δ = .09, p < .001, and substance use, β = .28, R²Δ = .08, p < .01. In addition, after-school supervision achieved significance in relation to boys’ delinquency, β = .13, R²Δ = .02, p < .05, and their substance use, β = .16, R²Δ = .02, p < .05.

Finally, to determine whether order of entry within blocks may have biased findings on individual predictors, analyses were rerun with father closeness forced to enter before mother closeness, and supervision preceding parental values. The only change in findings was that among boys, effects for father closeness came to approximate those for mother closeness in relation to delinquency, father: R²Δ = .07, p < .01; mother: R²Δ = .07, p < .01, and substance use, father: R²Δ = .04, p < .01; mother: R²Δ = .06, p < .01.

DISCUSSION

In conjunction with prior evidence obtained with affluent high school students, findings of this study indicate that (1) incidence of clinically significant depressive symptoms can be unusually high among suburban adolescent girls, (2) substance use is high among both male and female teenagers, (3) internalizing symptoms may often be implicated in substance use, and (4) peers may often actively approve of substance use among adolescent boys. Exploration of potential pathways to adjustment problems revealed that achievement pressures (internalized and from parents) can be implicated in disturbance across diverse domains, yet seem to confer no benefits in terms of facilitating children’s academic success. Also of apparent significance is isolation from adults—particularly levels of perceived closeness with mothers; and for girls, the presence of adult supervision in the hours immediately after school.

Replicating Prior Findings: Internalizing Problems and Substance Use among Affluent Youth

Viewed along with Luthar and D’Avanzo’s (1999) findings on the incidence of different problems among an independent cohort of suburban youth, results of the present study suggest that vulnerability may be pronounced during the adolescent years, and with some gender specificity of domains. Frequency of substance use was three times as high among seventh graders in this study, whose mean age was almost 13 years, as compared with sixth graders, whose mean age was about 12 years. Among girls in the seventh grade, rates of clinically significant depression were twice as high as those in normative samples of the same age. In contrast, among both boys and girls in the sixth grade, clinically significant depression rates approximated those in normative samples, and anxiety rates were markedly lower (by a factor of four).

Trends such as these may reflect a sharp increase in symptoms among suburban youth between the ages of 12 and 13 years, but equally could be an artifact of cohort differences. Various processes linked with the advent of adolescence might be implicated, such as hormonal changes and advances in formal operational thinking (e.g., with high preoccupation with personal identity). It is equally plausible, however, that the sixth-grade cohort in this study was, for whatever reason, a group of youngsters that was somewhat atypically well adjusted. Prospective research spanning the middle and high school years is necessary to establish whether subjective distress and substance use among suburban youth do, in fact, show (atypically) sharp increases at the onset of adolescence, and, as suggested by prior research (Luthar & D’Avanzo, 1999), remain high through the remaining high school years as well.

The consistency in findings on depression among 13-year-old girls in this study, with data on their 16-year-old counterparts in the prior study (Luthar & D’Avanzo, 1999), suggests the value of considering stressors that might be unique to these girls, and possibilities include aspects of gender-role socialization and self-image. In upper-middle-class communities, parents’ standards for academic excellence are typically high for adolescent daughters as well as for sons. Yet, clinical and qualitative reports suggest that the former are far more likely to contend with conflicting messages from the peer group and from the media with regard to displays of academic competence being “nonfeminine” and thus undesirable (e.g., Gjerde, 1995; Nolen Hoeksema, 1990; Pipher, 1996; Proweller, 1998). Additional stress may derive from high concern with personal appearance (Pipher, 1994), as dissatisfaction with their physical appearance tends to be particularly pronounced among young European American women (Abrams, Allen, & Gray, 1993; Cash & Henry, 1995; Rucker & Cash, 1992).

With regard to salient correlates of substance use, findings of this study are consistent with prior suggestions that affluent youth may often use substances in efforts to alleviate personal distress (Luthar & D’Avanzo, 1999; Way et al., 1994). Substance use showed significant, unique associations with four of the five self-reported maladjustment indices among girls: physiological anxiety, social anxiety, depression, and delinquent behaviors. Among boys, effects were obtained for physiological anxiety, depression, and delinquency; these three collectively accounted for over half of the variability in substance use scores.

The present findings also support suggestions of another risk factor previously found among suburban adolescent boys—peer approval of their substance use. Regression analyses with the present sample, like those of Luthar and D’Avanzo (1999), revealed significant links between boys’ substance use and peer ratings of their popularity. Furthermore, simple correlations in this study showed that the substance use–popularity link was significant only among seventh-grade boys (not among sixth graders); among these older boys, popularity also showed modest correlations with self-reported delinquency. In tandem, these results lend some support to Moffitt’s (1993) thesis regarding the growing endorsement of counterconventional behavior patterns with the onset of the adolescent years.

The present findings on boys’ substance use and popularity are consistent with reports that substance use among adolescent males, more so than females, is tied to social conformity motives such as drinking to fit in with a group of peers (Cooper, 1994). Longitudinal research by Feldman, Rosenthal, Brown, and Canning (1995) showed that middle-school boys who were best liked by their peers came to be among the most gregarious in high school, with gregariousness involving “partying” and heavy drinking. Similarly, analyses of trends in higher education indicate that contemporary American college students, particularly at the more renowned universities, are caught up in a triad of alcohol, spectator sports, and partying (Sperber, 2000), and media reports have drawn attention to serious levels of drinking in several elite college fraternities (e.g., McCormick, 2000).

Among both boys and girls in this study, links between substance use and peer rejection differed from those previously documented among high school students (Luthar & D’Avanzo, 1999). In this middle school sample, boys who used substances tended to be of “controversial” (Coie & Dodge, 1988) status, with frequent positive and negative nominations from peers, whereas among high school boys, substance use was related only to peer acceptance. Among girls, by contrast, substance use was unrelated to both positive and negative peer ratings among these middle school students, whereas it had been linked with both among high school girls. The dissonance in findings may reflect differing views of the deviance of substance use, depending on age and gender. A 16-year-old substance-using boy is less likely to elicit peers’ disapproval than is his 11-year-old counterpart. Similarly, substance use would probably have lower visibility among younger than older girls, given differences in actual use, as well as willingness to publicly acknowledge behaviors typically denounced among females (Luthar & D’Avanzo, 1999).

Collectively, indices of academic competence showed strong negative links with substance use among middle school students in this study, as they had among suburban tenth graders (Luthar & D’Avanzo, 1999). In this sample, unique links were found with poor school grades among both boys and girls: with teacher-rated problem behaviors among boys, and with absenteeism among girls. Findings such as these may reflect academic difficulties that arise from substance use, as a result for example, of attenuated attention and concentration or premature involvement in nonacademic activities (Baumrind & Moselle, 1985; Newcomb, 1987; Newcomb & Bentler, 1988). Equally, however, they might reflect the tendencies of academically failing children to disengage from school and to gravitate toward deviant activities, including substance use (Lynam, Moffitt, & Stouthamer-Loeber, 1993).

Pathways to Maladjustment: Achievement Pressures

Exploratory analyses of causes for suburban children’s distress yielded results consistent with expectations on the role of achievement pressures. Students’ levels of maladaptive perfectionism had robust associations with their subjectively perceived distress, and also showed significant links (although lower in magnitude) with their self-reported delinquency. Notably, these findings are unlikely to simply reflect artifacts of shared measurement in self-reports, because effects were statistically significant even after partialing out effects of several potentially overlapping indicators, each also based in self-report. Furthermore, there is little conceptual overlap in items across the instruments concerned. Two of the four mal-adaptive perfectionism subscales pertained to parents’ expectations, and the other two pertained to children’s doubts and mistakes specific to achievement tasks, as opposed to their generalized depressive affect or free-floating anxiety. In sum, then, results of this study do suggest that suburban youngsters who set excessively high standards for themselves in relation to personal accomplishments are likely to experience high disturbance across both internalizing and externalizing domains.

Although the present findings on perceived parental achievement values were more equivocal, we believe that this construct merits further attention in developmental research. Perceptions of parents’ emphasis on children’s achievements, relative to their personal well-being, showed links with high distress only among girls in this study. At the same time, it was linked with low rather than high levels of substance use among girls. Contrasting with this modest evidence for the importance of this construct are indications of its high significance both in prior clinical writings (Miller, 1997), and in research with adults, which indicate high depression and anxiety among those who emphasize extrinsic goals such as wealth, relative to intrinsic goals such as affiliation (Kasser & Ryan, 1993, 1996). In future developmental research, it would be useful to examine whether refinements of the measure used in this study (e.g., a shortened version that retains only the most unequivocal items; see DeCarlo & Luthar, 2000) might yield stronger support for hypothesized links. Also useful would be further examination of whether suburban adolescent girls may react more than do boys to parental overemphasis on achievements, and whether this construct might exacerbate vulnerability in some areas (distress), even as it serves protective functions in others (substance use).

Whereas the two indices of achievement pressures, considered together, showed various links with students’ self-reported maladjustment, it is noteworthy that neither was associated with their academic grades. Conjointly, these findings suggest that a disproportionate emphasis on children’s achievement, not uncommon in upwardly mobile, suburban communities, not only has the potential to engender distress among children but also has real constraints in terms of the capacity to generate the successes so pervasively exhorted.

Isolation from Adults

Among the strongest associations found in this study were those that involved closeness with mothers. This construct was significantly linked with distress and substance use among both boys and girls, and in addition, with boy’s delinquency levels, accounting for 10% of variance or more in each case. Furthermore, across multiple indices and among both boys and girls, links that involved closeness to mothers remained significant, even when closeness to fathers was given priority in the regression equations. Conversely, with mothers considered first, closeness to fathers showed few associations with adjustment outcomes.

A stark exception to this general pattern was that closeness with fathers had significant links with girls’ academic grades, whereas there were no effects for closeness to mothers. Although this lone association may be treated as a random aberration, its effect size compared well with many others in this study (accounting for 10% of variance), and it is consonant, furthermore, with arguments about gender-role socialization (Eccles, Wigfield, Harold, & Blumenfeld, 1993; Frome & Eccles, 1998). As noted earlier, preadolescent and adolescent girls are believed to struggle with peer group sentiments that to succeed academically is somehow unfeminine (e.g., Gilligan, 1982; Goodenow, 1993; Steele, 1997). Thus, the degree to which these girls feel close to (and thus potentially identify with) their fathers—individuals who often model goal-directed, achievement-oriented behaviors—could be highly significant in fostering their own positive attitudes toward academic achievements.

Two caveats must be noted in considering the overall pattern of findings obtained for closeness to mothers versus fathers. First, the results do not indicate greater felt closeness to mothers as opposed to fathers (in fact, boys’ mean scores were higher for the latter); rather, the data simply suggest that perceived relationships with mothers may have more pronounced ramifications for suburban children’s self-reported maladjustment. Second, it is entirely possible that fine-grained measurement of parent–child relationships would lead to a different set of conclusions than those indicated by the present analyses. The focus in this study was on relatively global constructs that encompass multiple aspects of relationships with each parent, derived to avoid Type I errors in statistical analyses. Some dimensions of relationships with suburban fathers, such as the degree to which they share leisure-time activities with their children, may well have particularly pronounced associations with children’s adjustment indices.

Overall, girls in this study were less likely than were boys to be unsupervised by adults after school, and lack of after-school supervision seemed to have greater ramifications for them. Regardless of order of entry of variables in regression equations, lack of after-school supervision had unique links with distress, delinquency, and substance use among girls, whereas among boys, it was linked with substance use only when given priority in the regression equations. Considering the gestalt of these findings along with those previously discussed, it is clear that the adjustment status of girls covaried with diverse indicators pertaining to adults in their lives—that is, closeness to mothers and fathers, perceived parent values, and the presence of after-school supervision. Among boys, perceptions of mothers was the only one of these factors to show unique links with maladjustment. In future prospective research, it will be important to ascertain whether suburban adolescent girls might in fact be sensitive to a wider range of family functioning indices than are boys, as is suggested in theoretical views on the significance of relationships in feminine development (see Gilligan, 1982; Maccoby, 1998).

Limitations

Salient among the limitations of this research is the cross-sectional measurement of all constructs, a design feature that precludes any conclusions about causality. It is possible, for example, that maladaptive perfectionism engenders depression; however, some children may become quixotically preoccupied with extraordinary achievements in efforts to assuage chronic unhappiness. By the same token, isolation from adults can certainly exacerbate distress levels, but troubled children are also more inclined than others to view their lives in a negative light.

Although heavy reliance on self-reports is often viewed as a research limitation, the decisions in this regard were made advisedly, because they were tied to the central objectives of the present study. A primary goal was to replicate findings of unusually high subjectively experienced disturbance among affluent youngsters, using a sample that was socioeconomically similar, but developmentally and geographically distinct, from that in the previous study (Luthar & D’Avanzo, 1999). Self-reports were also of critical relevance to the second objective; that is, to explore the degree to which children’s appraisals about particular aspects of their everyday lives might contribute most substantively to their generalized malaise, or proclivity to counterconventional behaviors, in this ostensibly privileged setting.

In relation to the second objective, the absence of a comparison group clearly precludes conclusions about whether the present findings apply to affluent youth any more than to others. As noted at the outset of this article, the socializing influences examined were selected on the basis of their likely salience in upper-class suburbia, yet many, clearly, will affect children more generally.

A final concern that could be raised about this investigation pertains to the “legitimacy” of children’s self-reported problems. It might be argued that suburban youngsters’ reports of high depression and anxiety may be an artifact of the ethos in their communities, where expectations of adolescent angst are commonplace. Although this contention is reasonable, less tenable would be the corollary argument that because of variations in their roots, children’s subjective experiences of unhappiness should be dismissed as inconsequential. Longitudinal research will be critical in illuminating this issue, for example, identifying in cross-lagged analyses the degree to which high self-reported distress among suburban youth does in fact presage subsequent deterioration in critical domains of functioning.

Resonance with Other Evidence: Implications for Action

To underscore the need for such rigorous developmental research in the future, we briefly summarize evidence from recent media reports, entirely convergent with our own preliminary findings. An article based on dozens of interviews with suburban youngsters indicated casual attitudes to substance use as early as the age of 13. “Boys … casually discussed the beers they liked and joked about (the) ‘international 4:20 club,’ which is supposed to be the universal time to smoke pot” (Franks, 2000, p. 102). Factors implicated included parents’ demanding careers and pressures faced by the children themselves: “We work so hard during the week, because of college pressure, that by the weekends we’re totally, like, Let the games begin” (Franks, p. 104). Another report based on interviews with psychotherapists underscored stressors deriving from a surfeit of extracurricular activities in upper-class communities. “Doctors claim that some children feel so much pressure for high performance that they develop stress-related symptoms like insomnia, stomachaches, headaches, anxiety, and depression” (Gilbert, 1999, p. F7). Several commentaries have drawn attention to the erosion, as a result of suburban children’s overscheduling, of “old-fashioned family togetherness” (Kantrowitz, 2000), and the “stabilizing, character-shaping experiences like suppertime conversations and family outings” (Belluck, 2000; see also Kantrowitz, 2000).

Accumulating evidence of such pressures suggests the value of considering some form of interventions for this subgroup of youth. Although it is widely believed that affluent youngsters who are troubled will receive appropriate help (e.g., psychotherapy), assumptions such as these are ill founded, for parents and teachers often do not seek help for children, even when aware that they are troubled (Puura et al., 1998). Furthermore, many upper-SES parents themselves struggle with the pressures of presenting as imperturbably strong, successful, and self-reliant, and are uneasy about seeking help because “they are supposed to be better able to handle their problems than [others]” (Wolfe & Fodor, 1996, p. 80). Notwithstanding the abundance of material resources, therefore, parents in suburban communities—typified by high anonymity and social isolation of residents (Buss, 2000)—may often be in considerable need of support both for themselves as individuals, and in relation to the challenges of parenting.

For those concerned with children’s development, there are several implications stemming from this work. For clinicians, educators, and policy makers, derivative messages are that (1) it is unwise to assume blithely that any subgroup of youth is invulnerable, and (2) it is critical to recognize explicitly that children and adolescents—including the most affluent—are not independent agents, able to secure their own help when experiencing high distress. For developmental researchers, these findings underscore the need to pursue further this line of enquiry. In the interest of both filling gaps in scientific knowledge and attending to social responsibilities (Fabes, Lynn Martin, Hanish, & Updegraff, 2000; McCall & Groark, 2000; Schwebel, Plumert, & Pick, 2000), scientists must continue the systematic investigation of problems of which there is now more than a fleeting glimpse: a series of pressures that may be relatively specific to upper-class suburbia.