|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
COPING WITH SOCIAL STRESS: IMPLICATIONS FOR PSYCHOPATHOLOGY IN
LISA M. SONTAG
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Lisa M. Sontag
To those who have always supported me and encouraged me to reach beyond my limits.
I thank my supervisory committee chair (Dr. Julia A. Graber) and members of supervisory
committee members (Dr. Eric Storch and Dr. Manfred Diehl) for their constructive and
thoughtful input. I thank Dr. Elizabeth Hayes for assisting me with my analyses. I thank the
participants of the study. I also thank the National Institute of Child Health and Human
Development and the National Institute of Mental Health for providing support for this proj ect.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ...............4.....
LIST OF TABLES ................ ...............7............ ....
LIST OF FIGURES .............. ...............8.....
AB S TRAC T ......_ ................. ............_........9
1 INTRODUCTION ................. ...............11.......... ......
Background ................. ...............11.......... ......
Peer stress ................. ...............13.................
Pubertal timing .................. ...............14.......... .....
Biological correlates of stress .............. ...............16....
Coping and Responses to Stress ................. ...............18........... ...
Proposed Study .............. ...............20....
Research question 1............... ...............21...
Research question 2............... ...............21...
2 METHODS .............. ...............23....
Participants .............. ...............23....
Procedure .............. ...............23....
M measures ................... .......... ...............24.......
Demographic Data............... ...............24..
Biological Response to Stress .............. ...............25....
Pubertal Timing ................. ...............26.......... .....
Experiences of Peer Stress............... ...............27.
Apprai sal ................. ...............27.................
Coping ................ ...............27...
Internal izing Distress............... ...............28
Ag gre ssi on ................ ...............29................
Analysis Plan .............. ...............29....
3 RE SULT S .............. ...............32....
Attrition Analy ses............... ...............32
Descriptive Analyses .............. ...............32....
Tests of Main Effects ................. ...............34................
Biological Response to Stress .............. ...............34....
Pubertal Timing ................. ...............36.......... .....
Peer Stress .............. ...............36....
Tests of M edition ................ ...............36........... ....
Pubertal Timing ........._.___..... .__ ...............36....
Peer Stress .............. ...............38....
4 DI SCUS SSION ........._.___..... .___ ...............5 1....
A HOME DATA COLLECTION PROCEDURE AND SALIVA SAMPLE
COLLECTION .............. ...............57....
B MIXED EFFECTS MODELS EXAMINING THE IMPACT OF TIME AND
SYMPTOMS OF PSYCHOPATHOLOGY ON PATTERNS OF CORTISOL
PRODUC TION ................. ...............58.................
LIST OF REFERENCES ................. ...............60................
BIOGRAPHICAL SKETCH .............. ...............65....
LIST OF TABLES
3-1 Means and standard deviations of adolescents' demographic information, health,
stress, coping, and psychological symptoms .............. ...............40....
3-2 Intercorrelations among demographic information, health, stress, coping, and
psychological symptoms ........... ......._._ ...............41....
3-3 Mixed effects models examining the association of internalizing distress and cortisol....42
3-4 Mixed effects models examining the association of aggression and cortisol ....................43
3-5 Pubertal timing predicting internalizing distress .............. ...............44....
3-6 Pubertal timing predicting aggression .............. ...............45....
3-7 Peer stress predicting internalizing distress ................ ...............46...............
3-8 Peer stress predicting aggression .............. ...............47....
B-1 Impact of time and internalizing distress on patterns of cortisol ................ ................. .59
B-2 Impact of time and aggression on patterns of cortisol ................ ......... ................59
LIST OF FIGURES
1-1 Bio-psychosocial framework ................. ...............22................
3-1 Path model for pubertal timing to internalizing distress ................. ................. ......48
3-2 Path model for peer stress to internalizing distress. ......_................. ................ ..49
3-3 Path model for peer stress to aggression. ....__ ......_____ .......__ ..........5
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
COPING WITH SOCIAL STRESS: IMPLICATIONS FOR PSYCHOPATHOLOGY IN
Lisa M. Sontag
Chair: Julia A. Graber
Major Department: Psychology
This study investigated the impact of social stress on girls' psychosocial development
during adolescence by integrating models of peer stress, pubertal timing, adrenocortical activity,
and coping. Based on a sample of 111 girls (Mage=11.84, SD = .77), we explored whether
patterns of cortisol, pubertal timing, and peer stress were associated with symptoms of
internalizing distress and aggression. Additionally, we examined whether coping strategies and
involuntary responses to stress mediated these associations. Cortisol reactivity was not
associated with adjustment; however, pubertal timing and peer stress were. Coping and
involuntary responses mediated the association between pubertal timing and peer stress and
adjustment. Findings indicate that learning to effectively adapt to stress may help girls transition
more successfully into adolescence.
Adolescence is a period of rapid growth and change, both physically and psychologically.
During the transition into adolescence, individuals must learn to manage the physical changes
their bodies endure due to the onset of puberty, changes in social context, and changes in school
environment (Graber & Brooks-Gunn, 1996). Such experiences inherently place young
adolescents at heightened risk for experiencing a great number of new daily stressors
(Brooks-Gunn, 1991; Ge, Conger, & Elder, 2001). Because the chance of encountering some
type of environmental or social stressor is inevitable, young adolescents must learn to adapt and
to cope with these life experiences in order to successfully navigate the transition into
adolescence and eventually into early adulthood.
During the transition into adolescence, individuals experience more stressful events
compared to older adolescents or children, which in turn is associated with an increase in such
outcomes as depressive affect (Graber & Brooks-Gunn, 1996; Rudolph, 2002). Of these stressful
events, off-time pubertal maturation and interpersonal stress within the context of peers pose
particular threats to adolescents (Graber, Brooks-Gunn, & Archibald, 2005). According to
Rudolph' s model of interpersonal stress (2002), girls are at a particular risk for experiencing
higher levels of peer stress and consequently psychological maladjustment compared to boys,
due to the fact that peer relationships, specifically with close friends, are of greater importance to
girls. In addition, some individuals exhibit disregulations in the production of cortisol in
response to stressors which increases the risk for poor psychological outcomes. For example,
variations in both basal activity and stress reactivity have been linked to internalizing and
externalizing problems among young adolescents (Klimes-Dougan, Hastings, Granger, Usher, &
Zahn-Waxler, 2001). However, exposure to potentially stressful situations is not always linked to
poor outcomes. The ability to cope with or respond effectively to stress relies heavily upon the
ability to interpret and react appropriately to the stressor itself or to the physiological, emotional,
and behavioral responses to the stressor. By integrating models of social and biological
indicators of stress and by using a more comprehensive view of how adolescents respond to
stressful experiences, this study explored what factors place some young adolescent girls at
greater risk for maladjustment in the face of stressful life experiences.
Stress may be best conceptualized as a multidimensional concept that includes not only
the stressor, but also the processing system, which includes the cognitive assessment of the
stimuli, and the stress response (Compas, Orosan, & Grant 1993). Although conceptually stress
may be a multilevel process, more often than not the experience of stress is quite automatic and
usually occurs unconsciously. Moreover, not all experiences are equally stressful to individuals,
thus making appraisals of potentially stressful experiences relevant (Lazarus & Folkman, 1984).
The transition into adolescence is marked not only by physical and sexual maturation, but
also by increased exposure to new, many of them challenging, events (Brooks-Gunn, 1991; Ge et
al, 2001; Graber & Brooks-Gunn, 1996). Although there is evidence that stress increases during
early adolescence, the experience of particular events as stressful is not necessarily pervasive
among all adolescents (Brooks-Gunn, 1991). Rather, the experience of an event as stressful is
likely due to interactions among various changing life events (e.g., interactions among hormonal
and physical changes, school transitions, peer group changes, and altering family dynamics) that
influence an adolescent' s interpretation of particular events as stressful (Brooks-Gunn, 1991).
One aspect of stress that seems to be particularly salient to adolescents is social stress.
Social stress represents stimuli within the context of social interactions or relationships that
evoke emotional, physical, and psychological arousal in individuals. Social stress may be
broadly characterized as experiences within a social context, including interactions with others
(e.g., peer stress), as well as how one's own development (e.g., puberty) is experienced in a
Peer stress may be best conceptualized as experiences within the peer context that exert
significant pressure on an individual's ability to function adaptively. More concrete examples of
peer stress include conflicts with close friends, being teased or hassled by other children, and
being rejected by other children. Peer stress becomes increasingly salient as a threat to
psychological well-being during adolescence due to the increasing amount of time spent with
peers relative to other interpersonal relationships (La Greca, 2001). The relative salience of peer
relationships and the impact of peer stress during the transition into adolescence are further
supported by findings from Rudolph (2002) that demonstrated that levels of stress increase from
preadolescence to adolescence independent of gender.
Increased exposure to interpersonal stress during childhood and adolescence has been
linked to disorders and symptoms of psychopathology, namely anxiety, depression, behavior
problems, and substance use (Coie, Dodge, & Kupersmidt, 1990; Hawker & Boulton, 2000).
Specific aspects of peer stress, peer rej section and victimization, have been consistently linked to
externalizing and internalizing problems in adolescents (Deater-Deckard, 2001). A study
examining aggression and peer victimization as predictors of behavioral and emotional
adjustment found that greater experiences of peer aggression predicted greater levels of both
self-reported aggression and delinquency (Khatri, Kupersmidt, & Patterson, 2000). Another
study, conducted by Storch, Nock, Masia-Warner, and Barlas (2003), found that among Hispanic
and African American children, overt peer victimization was positively associated with
depressive symptoms and aspects of social anxiety. Additionally, relational victimization was
uniquely associated with depressive and anxious symptoms for girls only. Recent work
examining cross-national consistency in the relationship between being bullied and psychosocial
adjustment found that being the recipient of bullying behaviors was linked to poor emotional
adjustment, poor school adjustment and poor relationships with classmates (Nansel et al., 2004).
As indicated, girls experience higher levels of peer stress compared to boys during
adolescence (Rudolph, 2002). This differential trend is likely due to the fact that peer
relationships, specifically with close friends, are of greater importance to girls. It is not to say
that boys do not have experiences that could be interpreted as stressful, but rather girls interpret
comparable situations as more stressful and detrimental to their interpersonal relationships.
One aspect of social stress that is particularly salient to adolescents is the experience of
pubertal changes and whether those changes reflect normative or off-time developmental trends
within one's peer group. In order to better understand the impact of pubertal development on
adolescents' overall experiences of stress and the relationship between puberty and
psychopathology, it is useful to distinguish between pubertal status and pubertal timing.
According to Graber, Peterson, and Brooks-Gunn (1996), pubertal status refers to the physical
and physiological maturational levels of an individual at a given point in time, relative to the
overall pubertal process. Pubertal timing, however, may be conceptualized as a social
construction and best represents whether an individual's maturation is earlier, on-time, or later in
comparison to same-age peers.
Research examining the relationship between pubertal maturation and changes in
depressive or aggressive affect has found few links between pubertal status and negative
outcomes (Brooks-Gunn, Graber, & Paikoff, 1994; Graber et al., 2005). In contrast, research
focusing on the effects of pubertal timing has demonstrated that the stress and conflict elicited by
one's off-time maturation is more strongly linked to psychological and behavioral maladjustment
(Graber et al., 2005 for a recent review). Additionally, this research has also indicated that early
maturing girls are particularly at risk for a variety of disorders or symptoms of disorders,
including anxiety, depression, aggression, substance abuse, and eating disorders. A study by
Graber, Seeley, Brooks-Gunn, and Lewinsohn (2004) examining the association between
pubertal timing and psychopathology in young adulthood found that women who were early
maturers were at unique risk for persistent difficulty and development of psychopathology during
adolescence and young adulthood. Other studies examining the role of pubertal timing found that
early maturing girls were at high risk for depressive symptoms, body dissatisfaction, aggression,
and delinquency problems (Caspi, Lynam, Moffitt, & Silva, 1993; Ge, Brody, Conger, Simons,
& Murry, 2002; Ge et al., 2001; Ohring, Graber, & Brooks-Gunn, 2002).
In order to explain the connection between hormonal and pubertal changes and negative
affective and behavioral outcomes, Brooks-Gunn and colleagues (1994) proposed a model that
included the possibility that social events or context, as well as the adolescent's perceptions of
his or her pubertal processes, contribute to or function as mediators of hormonal effects on
negative affect and behavior. In testing this model, Brooks-Gunn and Warren (1989) found that
social events, not pubertal development, accounted for a good deal of variance in affect and
behaviors. In terms of social processes, adolescents who mature early or late in comparison to
their peers are placed in socially "deviant" categories compared with on-time peers (Petersen &
Crockett, 1985). Keeping in mind that girls as a group tend to mature a couple of years earlier
than boys, early maturing girls are deviant in comparison to both other girls and boys, which
places these girls at a heightened risk for maladjustment.
This study explored the effects of pubertal timing within a social processes framework in
order to expand on previous findings that have linked the social experience of being "off-time"
and particular emotional or behavioral changes. As highlighted by Brooks-Gunn and colleagues
(1994), social factors play a significant role in connecting the pubertal timing and negative affect
or behavior. The way an individual interprets his or her timing is likely to influence the types of
social interactions he or she experiences, and in turn is likely to influence potential affective or
behavioral changes. Research has shown that early maturing girls consistently report poor quality
of relationships at mid-adolescence and young adulthood (Graber et al., 2004). By examining
how early maturing girls cope with or adapt to potentially stressful situations within the peer
context, a better understanding of the mechanisms by which pubertal timing relates to
maladjustment may emerge.
Biological correlates of stress
Under conditions interpreted as stressful, the limbic-hypothalami c-pituitary-adrenal
(LHPA) axis in mammals (including humans) responds to this experience via a series of events
that result in the secretion of glucocorticoids (cortisol in human and nonhuman primates, and
corticosterone in rodents) from the adrenal cortex (Cicchetti & Walker, 2001). Prolonged
exposure to environmental and internal stress, and in turn disregulations of cortisol levels, can
affect psychological well-being. Past studies investigating the association between adrenocortical
activity, stress, and psychosocial well-being found that variation in both the basal activity and in
stress reactivity has been associated with psychopathology (Angold, 2003; Goodyer, Herbert,
Tamplin, & Altham, 2000; Klimes-Dougan, Hastings, Granger, Usher, & Zahn-Waxler, 2001;
Southwick, Vythilingam, & Chamney, 2005). However, much of this literature has focused
primarily on diurnal patterns of change and investigated cortisol reactivity in clinical populations
(Oosterlaan, Geurts, Knol, & Sergeant, 2005; Van Bokhoven et al., 2005; Van de Wiel, Van
Goozen, Matthys, Snoek, & Van Engeland, 2004). Therefore the question still remains as to how
patterns of cortisol reactivity in normative samples of adolescents reflect experiences of stress,
particularly social stress, and how these patterns of reactivity are associated with changes in
negative affect and behavior.
For individuals who follow a typical day-night schedule, cortisol production peaks during
the last few hours of sleep (Gunnar & Vazquez, 2001 for a review). Cortisol levels then decline
rapidly throughout the morning and decrease gradually during the remainder of the day. When
stress is induced, cortisol secretion initially increases. Following the onset of a stressor, cortisol
usually takes about 10-15 min to produce a rise in circulating cortisol levels. About 20-30 min
after the stressor, peak concentrations of cortisol are found in saliva. Following this peak, the
negative feedback loop of the LHPA axis begins to reestablish homeostatic functioning by
reducing the level of circulating cortisol.
Recent research examining the association between changes in daytime cortisol levels in
response to social challenges (i.e., public speaking task or mother-child conflict task) has shown
that adolescents and children tend to display increases in salivary cortisol approximately 20 to 30
min following exposure to a social challenge task, and then display a decrease in levels of
cortisol (Granger, Weisz, & Kauneckis, 1994; Klimes-Dougan et al., 2001; Schmidt et al., 1999).
In addition, findings from these studies have indicated that maintaining high levels of cortisol or
displaying strong task-related decreases are associated with higher levels of internalizing and
Findings such as those mentioned previously point to a link between stress response and
psychopathology; however, activation of the L-HPA axis may not be directly linked to
symptomatology. Additionally, there are few conclusive or consistent findings within the child
and adolescent research about whether a true relationship between cortisol levels and
psychopathology exists. Therefore, more research examining the association between cortisol
and psychopathology is needed, especially within the adolescent female population.
Coping and Responses to Stress
Because of the increase in stressful experiences during the transition to adolescence
(Brooks-Gunn, 1991), learning to adapt to the stressor and manage one's response to the stressful
situation is crucial to leading a healthy and productive life. However, the effects of stress vary by
individuals. Whereas most people fare quite well in the face of typically occurring stressful
events, other individuals manifest symptoms of psychopathology, as mentioned previously. But
the question therein lies, why do individuals demonstrate differential outcomes to stress exposure
and how do these differences emerge? Coping strategies offer one of the mechanisms through
which stress differentially affects individuals. Understanding the process of coping and how
adolescents effectively or ineffectively react to stressful stimuli allows researchers to understand
better why particular individuals adapt well or adapt poorly.
As suggested by Lazarus and Folkman (1984), coping involves efforts to manage specific
external and/or internal demands that are appraised as taxing or exceeding the resources of the
person. Coping differs from general responses to stress in that coping is viewed as volitional or
under the person's control. In addition, Lazarus and Folkman emphasize the role of cognitive
processes in determining what is experienced as stressful and how one copes.
Theoretical frameworks of coping, such as the one developed by Lazarus and Folkman
(1984), adequately target volitional responses to stress; however, few models have included
involuntary responses to stress. One comprehensive model, developed by Compas, Connor,
Saltzman, Thomsen, and Wadsworth (1999), includes both volitional and involuntary responses
to stressful stimuli. According to this model, stress response can be characterized as being either
effortful or involuntary. Effortful responses are cognitive and behavioral reactions to stress
experienced as volitional and purposeful such as acceptance, denial, and avoidance. Involuntary
responses, as defined by Compas et al. (1999), include temperamentally based and conditioned
cognitive, behavioral, and physiological reactions to stress that may or may not be within
conscious awareness and are not under personal control, such as physiological arousal, intrusive
thoughts, and rumination. In combination with the volitional responses to stress (i.e., coping),
involuntary responses create a more representative model of how individuals respond to external
and internal stressors throughout their lifespan.
Compas and colleagues (1999) further differentiated effortful and involuntary responses
to stress by distinguishing between engagement and disengagement responses. Engagement
responses are directed toward a stressor or one's reactions to the stressor and include approach
responses. Disengagement responses are directed away from these targets and include avoidance
responses. Given this model, an individual may respond to a stimulus in one or more of the
following manners: effortful engagement (primary and secondary control), effortful
disengagement, involuntary engagement, and involuntary disengagement. Connor-Smith et al.
(2000) tested this model using a sample of adolescents and their parents. Findings indicated that
not only was there adequate to excellent internal consistency for each subtype of response to
stress (i.e., primary control, secondary control, disengagement coping, involuntary engagement,
and involuntary disengagement) but subscales of coping, specifically, mapped onto other
standard measures of coping quite clearly. More important, however, was the finding that factors
within this model of responses to stress correlated with internalizing and externalizing behavior.
Thus, these exploratory analyses suggest that a model including both coping and involuntary
responses adequately captured how adolescents adapt to stressors. Additionally, how one copes
with stress was associated with emotional or behavioral outcomes.
Other studies examining the association between responses to stress and emotional and
behavioral outcomes have consistently demonstrated that style of response mediates the
relationship between stress, specifically family conflict and chronic pain, and anxiety,
depression, and aggression (Jaser et al., 2005; Wadsworth, Raviv, Compas, & Connor-Smith,
2005). Extrapolating to other realms of stress, it is likely that an individual's type of response to
social stress (e.g., peer stress) will mediate the relationship between social stress and
psychopathology during adolescence. However, little empirical work has examined the effects of
responses to stress on this relationship and even fewer studies have extended this research to
examining how coping may serve as a mechanism through which biological indicators of stress
(i.e., cortisol) and pubertal timing influence poor outcomes.
This study investigated the impact of stress reactivity and social factors on negative
outcomes from a bio-psychosocial perspective as a means to explain emotional and behavioral
changes that occur during the transition into adolescence. By integrating several theoretical
models that focus on the context of peer stress, pubertal timing, adrenocortical activity, and
coping (Figure 1), this study aimed to provide a more comprehensive approach to understanding
the pathways to adjustment. Within this framework, this study examined physiological and
environmental indicators of social stress and how these indicators were associated with
symptoms of psychopathology, in particular internalizing distress and aggression. Additionally,
this study explored whether responses to stress mediated this relationship. Specifically, the study
examined the following questions:
Research question 1
Are physiological and environmental indicators of social stress associated with symptoms
of internalizing distress and aggression?
Specifically, is biological stress reactivity (i.e., pattern of cortisol levels) associated with
symptoms of psychopathology? Because of the exploratory nature of the study with
respect to the Eindings associated with cortisol, it was predicted that pattern of reactivity
was associated with symptoms of internalizing distress and aggression. However,
specific predictions about the association between particular patterns of cortisol and
behavioral outcomes were not made.
Is pubertal timing associated with symptoms of intemalizing distress and aggression?
This study predicted that girls who reported early pubertal timing would demonstrate
higher levels of internalizing distress and aggression compared to girls who reported
being on-time or late.
Are experiences of peer stress associated with symptoms of internalizing distress and
aggression? It was hypothesized that girls who reported a greater number of stressful peer
experiences would display more symptoms of intemalizing distress and aggression.
Research question 2
Do coping strategies mediate the association between biological and environmental
indicators of stress and symptoms of psychopathology?
Do coping strategies mediate the association between stress reactivity (i.e. cortisol
patterns) and symptoms of internalizing distress and aggression? It was hypothesized that
the 5 coping strategies or responses to stress (i.e., primary control, secondary control,
effortful disengagement, involuntary engagement, and involuntary disengagement) would
mediate the association between stress reactivity and symptoms of psychopathology.
More specifically, it was predicted that effortful disengagement and involuntary
disengagement would serve as mediators between stress reactivity and internalizing
distress, whereas involuntary engagement would mediate the association between stress
reactivity and aggression. It was expected that primary and secondary control coping
strategies would mediate the association between stress reactivity and internalizing
distress and aggression.
* Do coping strategies mediate the association between pubertal timing and symptoms of
internalizing distress and aggression? It was expected that coping strategies would
mediate the association between pubertal timing (i.e., early vs. on-time or late) and
symptoms of psychopathology in the same way that was described for stress reactivity.
* Do coping strategies mediate the association between peer stress and symptoms of
internalizing distress and aggression? It was hypothesized that coping strategies would
mediate the association between peer stress and symptoms of internalizing distress and
aggression in the same way that was described for stress reactivity and pubertal timing.
Coping Internalizing Distress
Figure 1-1. Bio-psychosocial framework
Participants were 111 adolescent girls (Mage = 11.84, SD = .77), who were drawn from a
larger proj ect, the Girls Health and Development proj ect, that examined biological correlates of
girls' social and emotional health as they made the transition to puberty. Approximately 15% of
the girls were in 5th grade, 47% in 6th grade, 3 1% in 7th grade, and 2% in 8th grade. Of the 11 1
girls, 48% were Caucasian, 36% African American, and 16% Hispanic. Girls and their families
were originally recruited in 1995-1996 from public schools in ethnically integrated,
working-class and middle-class communities in the New York City area. The study included 138
girls at baseline; 1 11 of these girls participated in the final follow-up, approximately three years
later. Analyses for the current study were completed using only girls who participated in the final
Girls and their families were recruited via their schools. Each time families participated,
parents signed informed consent forms at their homes, on their own behalf and for their child's
participation. Girls gave assent for their participation. Mothers and children completed several
tasks (Appendix A) during a home visit and completed surveys during the 3 consecutive days
following the home visit. Girls completed surveys on their own affect and behaviors. Surveys
were picked up from the families' homes 3 days after the home visit. Girls received a gift (tote
bag) for their participation and mothers were paid $75.
Data concerning the characteristics of the participants were collected using standard items
concerning age, household structure, and race/ethnicity. Girls reported their date of birth, and
exact age was calculated from date of home visit. Mothers reported on family characteristics and
child race/ethnicity. Race/ethnicity was collapsed into three categories: White, African
American, and Hispanic. For regression analyses, two dummy coded variables were used:
African American versus others and Hispanic versus other, with White as the omitted group.
Family SES was assessed using the Hollingshead Scale (Hollingshead, 1975), a scale
based on survey data of social status tied to U.S. census categories. The standard scoring
protocols were used for different household types. Mothers reported on the education and
occupation of themselves and any other parent/caregiver in the household. For each head of
household, education (coded on a 7-point scale from 1 = less than 7th grade to 7 = graduate
professional) was multiplied by a weight of 3 and occupation (coded on a 9-point scale from
1 = farm labor and menial service workers to 9 = higher executives and major professionals) was
multiplied by a weight of 5. Thus the range of possible scores for family SES was 8 to 66
(Hollingshead, 1975). For one-parent households, the education score and occupation score were
summed to yield an SES score. In two-parent families, one parent in paid employment, the
family SES was derived by adding the mean of the parents' education scores to the occupation
score for the employed parent. If both parents were employed, the family SES score was the
mean of the parents' SES scores.
Biological Response to Stress
During the in-home visit, 5 saliva samples were collected to assess cortisol levels
throughout a battery of tasks. See the Appendix A for a complete list of tasks. Approximately
15-20 min occurred between collections. The unit of measurement for cortisol assayed was Cpg/dl.
Saliva was collected via Salivettes. Girls kept the cotton roll in their mouths for 2 min,
resulting in sufficient amounts of saliva for repeat assays. Girls ingested a small amount of
sweetened Kool-Aid to stimulate saliva flow. Stimulating saliva flow with orally administered
substances such as powdered drink mix was recommended protocol at the time this protocol was
implemented (Granger et al., 1994; Gunnar & Vazquez, 2001; Hertsgaard, Gunnar, Larson,
Brodersen, & Lehman, 1992; Nachmias, Gunnar, Mangelsdorf, Parritz, & Buss, 1996).
Subsequent to the development of this procedure, it was found that Kool-Aid may alter the
results of the assay of cortisol from saliva samples (Schwartz, McFadyen-Ketchum, Dodge,
Pettit, & Bates, 1998). However, these effects were found to vary by assay kit used. The kit used
in the present study was not highly susceptible to this problem. Saliva samples were stored in a
refrigerator before being transported to Columbia-Presbyterian Reproductive Endocrinology
Department in New York City, where they were stored in a freezer at -20oC until assayed for
cortisol. Cortisol concentrations were determined using a radioimmunoassay adapted for
measurements in saliva (Diagnostic Products Company). The lower detection limit for this assay
was 0.02 Clg/dl per 200 Cll of saliva. Saliva samples were centrifuged at 3000 rpm for 10 min. A
200 Cll sample was used for duplicate analysis. All samples from a participant were analyzed in
one assay run, with inter- and intra-assay coefficients of variations less than 5 and 3%,
Certain samples were excluded from analysis due to problems with the saliva sample. The
lab processing the samples excluded participants' samples if the samples were insufficient to
process. Log transformations were applied to the cortisol data to reduce the large positive skew,
which is consistent with standard procedures in cortisol research (Gunnar, 2000). However,
original units were reported for ease of interpretation.
Pubertal timing was based on self-reported age of menarche obtained during the home
visit. A recent review of the use of self-reported menarche noted that accuracy of recall is
reduced as the time interval between menarche and recall is increased (Coleman & Coleman,
2002). However, if assessed within a period of 1 year from menarche, girls can typically recall
menarche within 1 month of the exact date. Given that the study's participants have recently
experienced menarche, the use of self-report can be considered a valid measure.
Using self-reported age of menarche, girls were classified as either early-maturers or other
(on-time or late). Cut-off ages by race (White, African American, and Latina) were derived from
a large, normative study of puberty (Herman-Giddens et al., 1997). Based on mean ages of
menarche reported by race in the Herman-Giddens study, girls that reported age of menarche
1/ SD below the mean were categorized as early. All other girls were in the other group (on-time
or late). According to an investigation of age at menarche for girls in the United States, age of
menarche for Hispanic girls tends to fall between ages for non-Hispanic white girls and
non-Hispanic black girls (Chumlea et al., 2003). Because the Herman-Giddens study only
examined White and African American girls, estimates for early Latinas were calculated as '/ SD
below the mean menarcheal age of the sample. Cut off ages for early maturing girls were: 12.28
y for White, 11.55 y for African American, and 11.97 y for Latina. Late maturation could not be
determined due to the mean age of assessment (just under 12 y) because the mean age of
menarche for each race as determined by Herman-Giddens et al. (1997) was greater than
12 years. Thirty five percent of the participants were classified as early maturers. Approximately
28% of White participants, 33% of African American participants, and 40% of Latina
participants were early maturers.
Experiences of Peer Stress
Participants' experiences of peer stress were assessed as part of the survey administered in
the days after the home visit using a stressful events checklist that was included in a version of
the Responses to Stress Questionnaire that was adapted for this study (RSQ; Connor-Smith,
Compas, Wadsworth, Thomsen, & Saltzman, 2000). The RSQ was developed to assess
experiences of stressful events and responses to stress in adolescent populations. The checklist
provided information about how many stressful experiences the participants had encountered
since the start of the school year. The RSQ can be tailored to specific stressors; this form of the
RSQ was tailored to assess peer stress. For the stressful events checklist, participants endorsed 9
items. Sum scores were calculated, with higher scores representing the experience of more
stressful events. Peer stress, as discussed in the study, was operationalized as the number of
stressful events a participant experienced.
Appraisal of stressful experiences was also assessed using the RSQ. Participants responded
to a single item that assessed overall how stressful the problems endorsed in the stressful events
checklist were for them on a likert scale (1 = not at all to 4 = very). Higher scores indicated that
the participant felt the experiences to be more stressful.
Coping was also assessed using the RSQ. Adapted from the original RSQ for the purposes
of this study, the measure consisted of 3 8 items rated on a 4-point likert scale (1 = not at all to
4 = a lot). The RSQ contains 5 sub scales: primary effortful control, secondary effortful control,
effortful disengagement, involuntary engagement and involuntary disengagement. In this study,
the primary effortful control subscale contained 7 items that assessed responses to stress
categorized as emotional expression, emotional regulation, and problem solving (a = .80). The
secondary effortful control subscale contained 4 items that measured responses to stress
categorized as acceptance, positive thinking, and cognitive restructuring (a = .70). The effortful
disengagement subscale contained 9 items and assessed denial, avoidance, and wishful thinking
(a = .75). The involuntary engagement sub scale contained 9 items, which assessed impulsive
action, physiological arousal, and intrusive thoughts or rumination (a = .78). Finally, the
involuntary disengagement subscale contained 9 items that measured inaction, emotional and
cognitive numbing, and involuntary avoidance or fleeing (a = .76). High scores on each sub scale
indicated that participants were more likely to engage in these types of responses in the face of
stress. Raw scores were reported as sum scores and used for reliability analyses. However, in
order to control for overall responding bias, factor scores reflecting the proportion of total
responses (e.g., sum of scores on primary control items/sum of all items) were used in analyzing
the associations with internalizing distress and aggression as recommended by Connor-Smith
and colleagues (2000).
Internalizing distress was assessed during the home visit using the anxiety/depression
sub scale of the Youth Self Report (YSR/1 1-18) created by Achenbach (1991). The YSR for ages
11-18 is a self-report measure designed to measure competencies, adaptive functioning, and
problems in youth and adolescents ages 11-18. It can be completed by those having a fifth grade
reading level or administered orally. Participants were asked to report on behavior now or within
the past 6 months and rated items on a likert scale (0 = not true, 2 = very true or often true).
Scores for the anxiety/depression subscale were reported as sum scores. The anxiety/depression
sub scale contained 16 items, with a reliability of a = .83.
Aggression was also assessed using the YSR/11-18 (Achenbach, 1991). Participants rated
items on a likert scale (0 = not true, 2 = very true or often true). Participants were asked to report
on behavior now or within the past 6 months. Scores for the subscale of aggression were reported
as sum scores. The aggression sub scale contained 19 items, with a reliability of a = .86.
The goal of this study was to examine physiological and environmental indicators of social
stress and how these indicators were associated with symptoms ofpsychopathology, specifically
internalizing distress and aggression. Additionally, this study explored whether responses to
stress mediated this association. First, attrition analyses were performed to determine if the
sample used in this study differed on key variables from the original sample of participants
recruited. Descriptive analyses were performed on key variables.
To address the first set of hypotheses, main effects were explored via ANCOVAs and
hierarchical linear regression. As a means to explore the main effect of stress reactivity on
internalizing distress and aggression, a combination of linear regression models and mixed
models was used. Within the cortisol literature, researchers use multiple approaches to examine
the association between stress reactivity (i.e., change in cortisol levels) and outcomes of interest.
The computation of the area under the curve (AUC) is a frequently used method in
endocrinological research and the neuroscience to explore information that is contained in
repeated measurements over time (Pruessner, Kirschbaum, Meinlschmid, & Hellhammer, 2003).
In the case of this study, AUC was employed to investigate task-dependent changes in cortisol
level over the period of the home visit. AUCoon-isol was calculated from zero and log
transformations were applied to the AUCoon-isolvalue for each participant to reduce the large
positive skew. To explore the main effects of stress reactivity, AUCoortisol was entered as the
independent variable in a linear regression model.
Although frequently used in endocrinological research to capture patterns of change over
time, the area under the curve as calculated from zero may neglect to capture idiosyncrasies of
change that occur regardless of one's absolute levels of cortisol. Also, as indicated, calculating
area under the curve requires data at each time point. Thus, in order to further explore potential
associations between patterns of change in cortisol production and symptoms of internalizing
distress and aggression, mixed effects models were used. These models addressed the question
of whether one' s pattern of production of cortisol over the 5 home visit samples changed in
association with one's symptoms of psychopathology. In addition, mixed models do not require
participants to have data at each time point. Thus, for this study, it was possible to estimate the
desired associations with fewer than 5 cortisol samples, allowing for a greater total sample size.
Addressing the remaining components of the first set of research questions, the main effect
of pubertal timing on internalizing distress and aggression was explored via a combination of
ANCOVAs and a hierarchical linear regression model. The main effect of peer stress was
explored using only a linear regression analysis.
Mediation was tested using the guidelines delineated by Baron and Kenny (1986). The
mediators were primary control coping, secondary control coping, disengagement coping,
involuntary engagement responses, and involuntary disengagement responses. Specifically,
separate regression models tested the effects of the social stress predictors (peer stress, pubertal
timing) or biological correlates of stress (cortisol) on each mediator and both internalizing
distress and aggression. A full model was then tested with the social stress predictor and all the
mediators entered on the same step. For mediation to occur, the predictor must have a significant
association with the mediator and the outcome variable, and, in the full model, the mediator must
predict the outcome variable and the size and significance of the association between the
predictor and the outcome variable must be reduced. In all models, race and family SES was
controlled for; in the models with peer stress as the predictor, appraisal was also controlled for.
Researchers comparing methods to test mediation have suggested that Baron and Kenny's
(1986) method neglects to provide a direct estimate of the size of the indirect effect of the
independent variable on the dependent variable (MacKinnon, Lockwood, Hoffman, West, &
Sheets, 2002; Preacher & Hayes, 2004). In turn they suggest supporting research findings with a
product of coefficients test, typically the Sobel Test, which states if ab (the indirect effect) is not
significantly different from zero, there is no effect of mediation (Sobel, 1982). However,
according to Preacher and Hayes (2004) this method is also misleading because there is reason to
be suspicious of the use of the normal distribution for computing the p value for the Sobel test
because the sampling distribution of ab may not be normal. To address this issue, Preacher and
Hayes (2006) recommend bootstrapping the sampling distribution of ab and deriving a
confidence interval with the empirically derived bootstrapped sampling distribution. This
approach has been suggested as a way of circumventing the power problem introduced by
asymmetries and other forms of non-normality in the sampling distribution of ab (Preacher &
Hayes, 2004; Shrout & Bolger, 2002). Given this recommendation, a test of the indirect effect
was performed to support mediation effects consistent with Baron and Kenny's (1986)
Due to the loss of 19.5% of the sample of girls from baseline (n = 13 8) to the final
follow-up 3 years later (n = 1 11), univariate analyses of non-response bias were
conducted to determine if the remaining sample differed on key demographic factors,
pubertal development variables, and outcomes (depression, peer problems, internalizing
and externalizing behavior) assessed at baseline. Results showed that families did not
differ on variables used in this study.
Some participants were excluded from particular analyses because they did not
have values for some variables. For analyses involving cortisol samples, some girls were
excluded because their saliva samples were insufficient to assay cortisol, whereas others
participated via phone interview and survey if they moved, and thus did not have a
cortisol sample. For the exact number of reports for each variable see Table 3-1.
Descriptive information (i.e., M and SD) are shown in Table 3-1 and correlations
among the variables of interest are shown in Table 3-2.
Contrary to the hypotheses, stress reactivity (AUCoor-tisol) was not associated with
internalizing distress or aggression (r = .02, ns, and r = .12, ns, respectively). In fact, only
two variables demonstrated significant correlations with stress reactivity: pubertal timing
(r = .23, p < .05), and involuntary engagement (r = .27, p < .05).
Pubertal timing also demonstrated significant associations with the outcomes. For
the dichotomous pubertal timing variable (early vs. Other), ANCOVAs (controlling for
SES and race) were conducted, as the F-values, means, and standard deviations for the
effects may be more illuminating than correlations. As expected, girls who were early
maturers reported more internalizing distress, F (1, 104) = 4.83, p < .05, with an average
score of 6.63 (SD = 4.76) compared to other girls (M = 4.57, SD = 4.44), and more
aggressive symptoms, F (1, 104) = 4.54, p < .05, with an average score of 8.77
(SD = 6.39) compared to other girls (M = 6.29, SD = 4.96).
Peer stress was positively associated with both internalizing distress (r = .40,
p < .01) and aggression (r = .23, p < .05). Appraisal of the stressful experiences was
positively associated with both peer stress (r = .41, p < .01) and internalizing distress
(r = .28, < .05 .
As anticipated, the potential mediators of the association between biological and
environmental indicators of social stress and adjustment demonstrated consistent
associations with the outcomes of interest. Internalizing distress was positively associated
with voluntary disengagement, involuntary engagement, and involuntary disengagement
responses (Table 3-2). Additionally as predicted, internalizing distress was negatively
associated with primary control strategies (r = -.48, p < .01), meaning that girls who used
problem solving and emotional expression and regulation tactics to cope with peer stress
reported fewer symptoms of anxiety and depression. The use of secondary control
strategies, such as acceptance or cognitive restructuring, was also negatively associated
(r = -.45, p < .01) with internalizing distress. This suggested that adolescent girls who
attempted to manage stressful peer experiences by using positive self-talk or accepting
these stressors as something from which to learn experienced fewer anxious and
depressive symptoms than girls were less likely to use these strategies.
Similar to the findings for internalizing distress, aggression was positively
correlated with involuntary engagement responses, and involuntary disengagement
responses (Table 3-2). As hypothesized, primary control engagement coping and
secondary control engagement coping were negatively correlated with aggression. The
use of disengagement coping strategies, such as denial, avoidance, and wishful thinking,
was not correlated with aggression.
Tests of Main Effects
As indicated, in order to test whether or not physiological and environmental
indicators of social stress were associated with symptoms of internalizing distress and
aggression, a series of hierarchical regression analyses were performed. All analyses
controlled for race/ethnicity (African American versus other, Hispanic versus other,
White as omitted group) and family SES.
Biological Response to Stress
Controlling for race and family SES, AUCoon-isol was not significantly associated
with internalizing distress, F (4, 65) = .59, p = .67, or aggression, F (4, 65) = .33, p = .86,
meaning amount of total stress reactivity over the 5 points of cortisol assessment were not
associated with symptoms of intemalizing distress or aggression. Sample sizes for these
analyses were much lower than the main group of participants (N= 111). This occurred
because in order to calculate an AUCoon-isol score, participants needed to have 4 out of the
5 samples and were not missing sample 1 or sample 5, in addition to needing race, SES,
internalizing distress, and aggression scores.
An unrestricted covariance structure was used for each model and all continuous
between-person variables were centered on the grand mean. Because race and family SES
were shown not to be significant predictors of AUCoon-isol, they were not included in the
mixed effect models. Researchers claim that including covariates that do not exhibit
associations with the dependent variable can generate spurious significant associations in
models, and thus it is recommended to exclude covariates from the analyses that would
otherwise be included in typical regression and ANCOVA analyses (Rovine, von Eye, &
Wood, 1988; Weisberg, 1979). One set of analyses was performed for each set of
symptoms (i.e., internalizing distress and aggression), resulting in 2 sets of analyses.
However, because internalizing distress and aggression showed not to be significant
predictors of participants' patterns of change in cortisol, results from the 2 sets of
analyses will be discussed in combination.
In Model 1 (see Tables 3-3 and 3-4), the variance in cortisol was partitioned into
within- and between-person variance; 41.18% of the variance was due to within-person
variability and 58.82% was due to between-person variability. In Models 2 and 3, fixed
effects were added in a systematic fashion and pseudo R2 -Staltistics were calculated to
determine the amount of variance these models accounted for relative to the baseline
In Model 2, time was included as a level-1 predictor of cortisol. Results
demonstrated that time was a significant predictor of cortisol levels and accounted for
43% of the within-person variance. In Model 3, the effects of internalizing distress
(Table 3-3) or aggression (Table 3-4) were added. Both internalizing distress and
aggression were not significant predictors and did not account for any between-person
variance. In addition, the interactions between time and internalizing distress and time
and aggression were not significantly associated with cortisol levels. Thus, although
results showed that cortisol levels changed over the 5 in-home collections, symptoms of
psychopathology were not associated with how cortisol levels changed over time. These
results corroborate the findings from the area under the curve analyses.
To test whether or not pubertal timing was associated with symptoms of
internalizing distress and aggression, a series of hierarchical regression models was
performed (Table 3-5). Controlling for race/ethnicity and SES, pubertal timing was
significantly associated with internalizing distress, such that girls who were early
maturers demonstrated higher levels of internalizing distress compared to other girls.
Controlling for race/ethnicity and SES, pubertal timing was also significantly associated
with aggression, such that girls who reported early maturation demonstrated higher levels
of aggression compared to other girls. These findings were consistent with bivariate
correlations reported in Table 3-2.
As expected, results indicated that peer stress was significantly associated with
both internalizing distress and aggression (Tables 3-7 and 3-8). Controlling for race,
SES, and appraisal, higher levels of peer stress were associated with greater symptoms of
internalizing distress and aggression.
Tests of Mediation
As indicated, a series of regression models were performed to test effects of
mediation as specified by Baron and Kenny (1986), with subsequent examination of tests
of mediation as proposed by Preacher and Hayes (2006).
Figure 3-2-1 shows the model testing mediated pathways of pubertal timing (early
vs. Other) to internalizing distress. The standardized P coefficients are shown in the
figures. In all, 5 tests of mediation (one for each mediator) were performed for
internalizing distress. As can be seen in Figure 3-2-1, there was evidence for full
mediation of the effect of timing on internalizing distress via primary control and
involuntary engagement responses. As per specifications of Baron and Kenny (1986), the
coefficient of pubertal timing predicting internalizing distress was reduced and was no
longer significant when the mediators were included in the model (i.e., P for pubertal
timing to internalizing distress, when mediated by primary control, was reduced from .20,
p < .10, to .08, ns, and was reduced from .20, p < .10, to .11, ns, when mediated by
involuntary engagement). Results from Preacher and Hayes (2006) test of the indirect
effect found that the indirect effect (ab = .1 1) of pubertal timing on internalizing distress
through primary control was estimated to lie between .03 and .27 with 95% confidence.
Also, the indirect effect (ab = .10) of pubertal timing on internalizing distress through
involuntary engagement was estimated to lie between .03 and .20 with 95% confidence.
For both indirect effects, mediation was said to occur because the indirect effects (ab)
were significantly different from zero.
The tests of mediation demonstrated that early pubertal timing predicted less use of
primary control strategies which subsequently predicted fewer symptoms of internalizing
distress. Conversely, early timing predicted greater use of involuntary engagement
responses, which subsequently predicted greater symptoms of internalizing distress.
Pubertal timing was not associated with secondary control strategies, voluntary
disengagement strategies, and involuntary disengagement responses.
Potential mediation effects on the association between pubertal timing and
aggression were not examined. Although initial analyses demonstrated that pubertal
timing was significantly associated with aggression (Table 3-6), analyses for tests of
mediation demonstrated that pubertal timing did not demonstrate a significant association
with aggression (P = .15, p = .14). In the tests of mediation, listwise deletion was
utilized, thus excluding participants who did not have scores for all the necessary
variables needed to illustrate mediation (i.e., pubertal timing, internalizing distress,
coping scores, and all covariates). This difference in analysis between the general
regression and the tests of mediation accounts for the emergence of a significant
association between pubertal timing and aggression in the original regression analysis
and the non-signifieant association that emerged in the test of mediation.
Figure 3-2 shows the model testing mediated pathways of peer stress to
internalizing distress; and Figure 3-3 shows the model examining the association between
peer stress and aggression. As shown in Figure 3-2 and 4, peer stress was positively
associated with voluntary disengagement; however, because voluntary engagement
strategies were not associated with either internalizing distress or aggression, mediation
effects did not emerge. Peer stress was not associated with involuntary engagement and
Two mediation effects emerged for both internalizing distress and aggression. As
can be seen in Figure 3-2, there was evidence for partial mediation of the effect of peer
stress on internalizing distress via primary control and secondary control strategies. The
coeffieient of peer stress predicting internalizing distress was reduced and was no longer
significant when the mediators were included in the model (i.e., P for peer stress to
internalizing distress, when mediated by primary control, was reduced from .35, p < .01,
to .26, p < .05, and was reduced from .35, p < .01, to .25, p < .05, when mediated by
secondary control). Noteworthy was the fact that the association between peer stress and
primary control was only a trend. However, using Preacher and Hayes' (2006) test of the
indirect effect supported the Einding of partial mediation, demonstrating that ab was
significantly different from zero with 95% confidence for primary control, ab = .10, CI
(.02, .21), and for secondary control strategies, ab = .10, CI (.01, .23).
Figure 3-3 shows evidence for full and partial mediation of the effect of peer stress
on aggression via primary control and secondary control strategies, respectively. The
coefficient of peer stress predicting aggression was reduced and was no longer significant
when the mediators were included in the model (i.e., P for peer stress to aggression, when
mediated by primary control, was reduced from .25, p < .05, to .20, p < .10, and was
reduced from .25, p < .05, to .17, p = .13, when mediated by secondary control).
Supporting these Eindings, ab was significantly different from zero with 95% confidence
for primary control, ab = .06, CI (.01, .16), and for secondary control strategies, ab = .08,
CI (.01, .24). For tests of the indirect effect, ab was significantly different from zero,
which demonstrated mediation. This means that greater peer stress predicted less use of
primary and secondary control strategies which subsequently predicted fewer symptoms
of internalizing distress and aggression.
Table 3-1 Means and standard deviations of adolescents' demographic information,
health, stress, coping, and psychological symptoms
Variable N M
Age 111 11.84
SES 110 33.07 11
Cortisol sample 1 89 .17
Cortisol sample 2 84 .14
Cortisol sample 3 80 .12
Cortisol sample 4 71 .09
Cortisol sample 5 73 .10
Cortisol AUC 72 9.72 11
Age of menarche 46 1 1.26
Peer stress 98 2.60
Degree of hassle 96 2.16
Primary control 101 .23
Secondary control 100 .13
Effortful disengagement 101 .24
Involuntary engagement 100 .20
Involuntary disengagement 100 .19
Internalizing distress 109 5.21
Aggression 109 7.07
Note. Coping and stress responses are reported as proportional scores.
Table 3-2 Intercorrelations among demographic information, health, stress, coping, and psychological symptoms
1 2 3 4 5 6 7 8 9 10 11 12
2. Cortisol AUC -.02 --
3. Pubertal Timing -.06 .23* -
4. Peer Stress .05 -.05 .02 --
5. Degree of Hassle .16 .06 -.09 .41 --
6. Primary Control .11 -.13 -.22* -.23* -. 18t
7. Secondary Control -.03 .20 -.05 -.27* -.20t .48 --
8. Vol. Disengagement -.05 -.09 .07 .19 .04 -.59* -.26--
9. Inv. Engagement .03 .27* .21* .17 .26* -.44* -.57* -. 19t
10. Inv. Disengagement -.08 -.11 .06 .13 .10 -.62* -.64* .01 .35*
11. Internalizing Distress .10 .02 .21* .40* .28* -.48* -.45* .19t .40* .42 --
12. Aggression -.04 .12 .21* .23 .13 -.34 -.4 -.05 .46~ .43 .56~ ---
Note. p <.10; p<.05; ** p<.01
Table 3-3 Mixed effects models examining the association of internalizing distress and cortisol
Parameter Model 1 Model 2 Model 3
Intercept Too -1.12*** -0.96*** -0.96***
(0.04) (0.04) (0.04)
Internalizing Distress 70, -.002
Rate of change
Intercept 710 -0.09*** -0.09***
Time X Internalizing 7,, -0.003
Within-person residual eij 0.07*** 0.04*** 0.04***
(0.02) (0.004) (0.004)
In initial status PUo, 0.10*** 0.11*** 0.12***
(0.65) (0.02) (0.02)
In rate of change p 0.006*** 0.005***
Pseudo R2 and Goodness of Fit
R2 E (within person variance) 0.43 0.00
R20 (variance in initial
R21 (variance in rate of 01
Deviance (-2LL) 260.4 135.1 153.4
ADeviance (-2LL) 125.3*** -18.3
Note. *** p<.001
Table 3-4 Mixed effects models examining the association of aggression and cortisol
Parameter Model 1 Model 2 Model 3
Rate of change
Time X Aggression 7,3
Within-person residual eg
In initial status Uo
In rate of change p,
Pseudo R2 and Goodness of Fit
R2 E (within person variance)
R20 (variance in initial
R21 (variance in rate of
Note. *** p<.001
Table 3-5 Pubertal timing predicting internalizing distress
Variable R2 AR2 B SEE B
Note. p <.05
Table 3-6 Pubertal timing predicting aggression
Variable R2 AR2 B SEE B
Note. p <.05
.08 1.02 .01
.50 1.35 .04
.01 .03 .03
.64 .51 .14
.81 .26 .38**
Table 3-7 Peer stress predicting internalizing distress
Variable R2 AR2 B SEE B
Degree of Hassle
Degree of Hassle
Note. ** p <.01
Table 3-8 Peer stress predicting aggression
Variable R2 AR2 B SEE B
Step 1 .031
African American .72 1.26 .06
Latina -.15 1.66 -.01
Family SES -.03 .04 -.09
Degree of Hassle .75 .57 .14
Step 2 .082* .051*
African American .73 1.23 .06
Latina -.13 1.62 -.01
Family SES -.03 .03 -.09
Degree of Hassle .20 .61 .04
Peer Stress .68 .31 .25
Note. p <.05
a) .20t / .08 for primary control
b) .20t /.11 for involuntary engagement
Figure 3-1. Path model for pubertal timing to internalizing distress. Each pathway controls for
race and family SES. At the top of the model, the P value was calculated for the
pathway between pubertal timing and internalizing distress. On the left side of the
model, the P values were calculated separately for each pathway for pubertal timing
to each potential mediator. The P values on the right side of the model were
calculated with all variables (covariates, predictors, and mediators) entered
simultaneously into the model. (a) On the right side, change in P values for pathway
between pubertal timing and internalizing distress when primary control is
simultaneously entered; (a) On the right side, change in P values for pathway between
pubertal timing and internalizing distress when involuntary engagement is
simultaneously entered. "f p < .10, p < .05, ** p < .01, *** p < .001
a) .35 / .26* for primary control
b) .35 / .26* for secondary control
Figure 3-2. Path model for peer stress to internalizing distress. Each pathway controls for race,
family SES, and degree of hassle. At the top of the model, the P value was calculated
for the pathway between peer stress and internalizing distress. On the left side of the
model, the P values were calculated separately for each pathway for peer stress to
each potential mediator. The P values on the right side of the model were calculated
with all variables (covariates, predictors, and mediators) entered simultaneously into
the model. (a) On the right side, change in P values for pathway between peer stress
and internalizing distress when primary control is simultaneously entered; (b) On the
right side, change in P values for pathway between peer stress and internalizing
distress when secondary control is simultaneously entered. "f p < .10, p < .05,
** p <.01, *** p <.001
a) .25* /.19t for primary control
b) .25 / .17 for secondary control
Figure 3-3. Path model for peer stress to aggression. Each pathway controls for race, family SES,
and degree of hassle. At the top of the model, the P value was calculated for the
pathway between peer stress and aggression. On the left side of the model, the P
values were calculated separately for each pathway for peer stress to each potential
mediator. The P values on the right side of the model were calculated with all
variables (covariates, predictors, and mediators) entered simultaneously into the
model. (a) On the right side, change in P values for pathway between peer stress and
aggression when primary control is simultaneously entered; (b) On the right side,
change in P values for pathway between peer stress and aggression when secondary
control is simultaneously entered. "f p < .10, p < .05, ** p < .01, *** p < .001
Although much is known about stress and coping during adolescence, few studies have
examined the influence of coping strategies on the relationship between social stress and
adjustment. More specifically, relatively little is known as to how adolescents cope with being
"off-time" with respect to puberty, and how responses to being off-time influence psychosocial
outcomes (Compas et al., 1999; Connor-Smith et al., 2000; Ge et al., 2001; Jaser et al., 2005;
Klimes-Dougan et al., 2001). The goal of this study was to examine the association of biological
and psychosocial indicators of stress and psychopathology from an integrated bio-psychosocial
perspective and to expand upon the current understanding of how biological and social factors
work in combination to influence negative outcomes during adolescence.
This study examined the effect of peer stress on adjustment and why the impact of peer
stress on adjustment varies among young adolescent girls. Consistent with prior research, girls
who experienced more peer stress reported higher levels of internalizing distress and aggression
(Deater-Deckard, 2001; Nansel et al., 2004; Rudolph, 2002). Moreover, as expected, coping
strategies, specifically primary and secondary control strategies, mediated both of these
associations. Specifically, girls who experienced more peer stress used fewer primary and
secondary control strategies, which are typically considered to be adaptive and effective means
of coping, such as problem solving, positive thinking, emotional regulation, etc. This association,
in turn, accounted for greater levels of internalizing distress and aggression. Recently studies
examining the impact of social stress on adjustment during adolescence have begun to explore
how stress responses influence this relationship; however, most of these studies have focused on
the direct association between stress responses and adjustment or examined how coping and
stress responses moderated the association between stressful life events and adjustment
(Connor-Smith, et al., Hampel & Petermann, 2006). Building upon prior research, the present
study focused specifically on peer stress and demonstrated that how an adolescent responds to
peer stress mediates the association between stressful experiences and adjustment during
adolescence. These findings are important because they provide a more thorough understanding
of how stressful experiences with peers impact adolescents' psychosocial well-being.
This study also examined the effect of pubertal timing on adjustment. As anticipated, girls
who were early maturers, compared to those who were on-time or late, demonstrated higher
levels of internalizing distress and aggression. This finding is consistent with prior research that
has demonstrated that experiencing puberty earlier than one's peers may place girls at a
heightened risk for a variety of behavioral and emotional problems (Graber, 2003; Weichold,
Silbereisen, & Schmitt-Rodermund, 2003). However, based on findings from this study, this
association may be partially accounted for by the way girls deal with being early maturers. The
use of primary control strategies (i.e., emotional expression, emotional regulation, and problem
solving) has been shown to attenuate the effects of stressful experiences, whereas, the use of
involuntary engagement (i.e., impulsive action, physiological arousal, and intrusive thoughts or
rumination) has been shown to intensify the effects of stress on the individual (Compas, et al.,
1999; Wadsworth, Raviv, Compas, & Connor-Smith, 2005). In this study, these two responses to
stress mediated the association between early pubertal timing and internalizing distress.
Meaning, early maturers used fewer primary control strategies and more involuntary engagement
strategies, which in turn accounted for greater levels of internalizing distress.
Although main effects of pubertal timing were demonstrated for both internalizing distress
and aggression, these effects were reduced in the full regression (i.e., when stress response
factors were included). The main effect of pubertal timing on internalizing distress was reduced,
but remained statistically significant, thus allowing for mediation effects of stress responses to be
examined, as recommended by Baron and Kenny (1986). However, the main effect of pubertal
timing on aggression was reduced to the point that it was no longer significant, hindering tests of
mediation. According to Baron and Kenny (1986), mediation effects cannot be tested if a
significant effect between the independent variable pubertall timing) and the dependent variable
(aggression) does not exist. The drop in significance was likely due to the exclusion of
participants that did not have responses to stress scores. Five of these participants were early
maturers and 5 of them were on-time or late. Proportionally, there was a greater drop in the
number of early maturers included in the analyses. For this reason, it likely became more
difficult to demonstrate significance. Because the inability to test mediating effects of coping
and stress responses on the association between pubertal timing and aggression was likely due to
an attrition effect, future research should revisit this issue in order to demonstrate whether or not
using particular coping strategies benefits early maturers.
Although an abundance of research exploring the relationship between pubertal timing and
maladjustment exists, few studies have attempted to explore exactly how pubertal timing relates
to adjustment and how girls cope with the social and emotional ramifications of being earlier
than their peers (Brooks-Gunn et al., 1994; Caspi et al., 1993; Ge et al., 2002; Graber et al.,
2005). The Eindings from this study are important because they demonstrate that some of the
association between early timing and adjustment may be accounted for by how adolescents cope
with or react to the stress of being "off-time."
Although examining the influence of individual coping strategies and stress responses is
useful and informative, future research exploring coping and peer stress should also examine
coping styles and how they impact adjustment. Recent research exploring coping has suggested
that individuals have a tendency to respond to stress using a particular set of coping strategies
and involuntary responses rather than responding with one strategy (Reinhard, Wolff, &
Wadsworth, 2006). Research that focuses on one type of stress response may be missing the big
picture of how individuals deal with stress. Thus, it may be informative to understand how these
different combinations of strategies (i.e., coping styles) either attenuate or exaggerate the effects
of peer stress.
Contrary to expectation, this study failed to demonstrate an association between stress
reactivity and symptoms of internalizing distress and aggression; as a result, potential mediating
effects of coping strategies could not be explored. This study found no connection between how
cortisol levels changed over the course of a home visit (that included a variety of tasks) and
symptoms of psychopathology. Much of the current literature has found associations between
cortisol reactivity and psychopathology, but mostly in samples reporting borderline or clinical
levels of internalizing and externalizing behaviors (Angold, 2003; Klimes-Dougan, et al., 2001;
Oosterlaan, et al., 2005). In a study similar in design to the current study (i.e., normative
adolescent samples exposed to social challenge tasks), Klimes-Dougan and colleagues (2001)
demonstrated links between stress reactivity and internalizing symptoms, but were unable to find
clear associations between stress reactivity and aggression. It was only until adolescents
displayed clinical levels of aggression that a significant association with stress reactivity
emerged. Thus, because most of the girls in this study displayed low or at most marginally high
levels of internalizing distress and aggression, associations between stress reactivity and
symptoms of psychopathology may not have emerged. In turn, further research examining
adrenocortical activity in normative adolescent samples should be conducted in order to provide
a more thorough understanding of whether or not a reliable connection between stress reactivity
and symptoms of psychopathology exists.
Some of the inconsistency in findings may be due to analytic differences. Most studies
examining stress reactivity in child and adolescent samples utilize reactivity categories (i.e.,
increased dramatically, remained stable, decreased slightly, etc.) to explore stress reactivity and
adjustment (Granger, et al., 1994; Klimes-Dougan, et al., 2001; Schmidt et al., 1999). This study
utilized area under the curve and mixed modeling to explore patterns of change. There exists a
debate as to what method is most appropriate to utilize when exploring adrenocortical activity
and adjustment. Thus further research exploring the method of AUC and mixed modeling in
different age groups of both boys and girls is necessary to understand the implications of using
these methods of analysis.
This study has taken informative steps in understanding better how social stress influences
psychosocial development; however, some limitations of the study exist. The examination of
stress and coping was based solely on adolescent girls, rather than both boys and girls. Although
researchers have suggested that social stress may be more detrimental to adolescent girls'
psychosocial well-being (Rudolph, 2002), it would be beneficial to examine further if gender
differences during early adolescence emerge with respect to how girls and boys cope with social
stress, and whether particular coping strategies are more problematic for one gender versus the
other. In addition, the study focused on early adolescence, particularly 12 year olds, which
prevented exploration of developmental changes in coping and stress responses over the course
of adolescence and the impact of late pubertal timing. Additionally, due to the fact that data were
only collected at a single time point, the direction of associations between social stress, coping,
and psychopathology were inferred from theoretical perspectives and previous research
exploring stress, coping and adjustment. Thus, results should be interpreted with some caution.
For example, it is unclear if girls who utilize particular types of coping strategies such as primary
and secondary control are more socially skilled and experience less peer stress. A Einal limitation
was the fact that participants were drawn from a large metropolitan area, limiting generalization
to adolescents in suburban and rural areas.
Despite the limitations, results of this study provide useful information regarding the
identification of specific coping strategies and stress responses that mediate the relationship
between stress and internalizing distress, and aggression. These Eindings speak to skills that
could be taught in programs that emphasize how to effectively adapt to the social and
psychological demands of transitioning into adolescence. Although the present study focuses on
negative outcomes, potential also exists to understand better what factors promote positive
development (i.e., prosocial behavior, academic success, positive self-esteem) in the face of
stressful social situations. Understanding these factors and how they affect psychosocial
development may ultimately expand our current understanding of resilience in the face of stress.
HOME DATA COLLECTION PROCEDURE AND SALIVA SAMPLE COLLECTION
2. Specimen and questionnaire delivery details
3. Saliva #1
4. Maze task
5. Neighborhood/peer questions
6. Saliva #2
7. Cold pressor task #1
8. Cold pressor questions
9. Cold pressor task #2
10. Cold pressor questions
11. Disagreement Questions
12. Saliva #3
13. Daily routines/ nutrition/ sleep questions
14. Physical exam
15. Menstruation & social activities questions
16. Saliva #4
17. Peer issues interaction task
18. Disagreement interaction task
19. Saliva #5
MIXED EFFECTS MODELS EXAMINING THE IMPACT OF TIME AND SYMPTOMS OF
PSYCHOPATHOLOGY ON PATTERNS OF CORTISOL PRODUCTION
Table B l. Impact of time and internalizing distress on patterns of cortisol
Model Level-1 Model Level-2 Model Composite Model
A Cortisol, = Ri, + e, ~ io, = 700 + 90, Cortisol,, = 700 + (ez, + 903,)
B Cortisol, = Ri, + zzTime, + e, or~ = 700 + 90, Cortisol,, = 700 + 710Time,
%, =710 Ae, + 90,, + Au, Timer; )
C Cortisol, = RT0 + n,Timell + ell ioz = Too, +,v01(Anlx/Dep Anzx/Dep) + 0,I Cortisol at = 700 + Trolime v +
ri,1 = 70 + 7,,(Anx/Dep -Anx/Dep)+ #0, 70,[( Anx /Dnep Anx /Dnep)(Time ii )
(ez, +o #u,+ ,Time, )
Note. (A) unconditional means model (partitions variance into between and within variance); (B) unconditional growth model (do
cortisol levels vary over time); (C) examines effect of internalizing distress on initial status and rate of change of cortisol
Table B2. Impact of time and aggression on patterns of cortisol
Model Level-1 Model Level-2 Model Composite Model
A Cortisol, = Ri, + e, ~ io, = 700 + 90, Cortisol,, = 700 + (ez, + 903,)
B Cortisol, = or~ + zzTimeO + ez, or = 700 + 90, Cortisolet = 700 + y0Time0 +
=710+A,(eO + 90, + Au,Timer,
C CortisEol = or, + %,~Timelj + ej no2 = 700 + 70, (Agg Agg)+ +, Cortisol at = 700 + 7/,0Tim~e I
Z11 = 710 + 7,(Agg-Agg)+ #1, 70, [(Agg Agg )(Time 23)
eO u, + Po;,,1Timery)
Note. (A) unconditional means model (partitions variance into between and within variance); (B) unconditional growth model (do
cortisol levels vary over time); (C) examines effect of aggression on initial status and rate of change of cortisol
LIST OF REFERENCES
Achenbach, T. M. (1991). Manual for the Youth Self-Report and 1991 Profile. Department of
Psychiatry, University of Vermont, Burlington, VT.
Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster Analysis. Beverely Hills, CA: Sage
Angold, A. (2003). Adolescent depression, cortisol, and DHEA. Psychological M~edicine, 33,
Brooks-Gunn, J. (1991). How stressful is the transition to adolescence for girls? In M. E. Colten
& S. Gore (Eds.), Adolescent stress: Causes and consequences (pp. 131-149). New York:
Aldine de Gruyter.
Brooks-Gunn, J., Graber, J. A., & Paikoff, R. L. (1994). Studying links between hormones and
negative affect: Models and measures. Journal of Research on Adolescence, 4, 469-486.
Caspi, A., Lynam, D., Moffitt, T. E., & Silva, P. A. (1993). Unraveling girls' delinquency:
Biological, dispositional, and contextual contributions to adolescent misbehavior.
Developmental Psychology, 29, 19-30.
Chumlea, W. C., Schubert, C. M., Roche, A. F., Kulin, H. E., Lee, P. A., Himes, J. H., et al.
(2003). Age at menarche and racial comparisons in US girls. Pediatrics, 111, 110-113.
Cicchetti, D., & Walker, E. F. (2001). Stress and development: Biological and psychological
consequences. Development & Psychopathology, 13, 413-418.
Coie, J. D., Dodge, K. A., & Kupersmidt, J. B. (1990). The role of poor peer relationships in the
development of disorder. In S. R. Asher & J. D. Coie (Eds.), Peer rejection in childhood
(pp. 274-305). New York: Cambridge University Press.
Coleman, L. and Coleman, J. (2002). The measurement of puberty: A review. Journal of
Adolescence, 25, 535-550.
Compas, B. E., Connor, J. K., Saltzman, H., Harding Thomsen, A., & Wadsworth, M. (1999).
Getting specific about coping: Effortful and involuntary responses to stress in
development. In M. Lewis & D. Ramsay (Eds.), Southling and stress (pp. 229-256).
Mahwah, NJ: Lawrence Erlbaum Associates.
Connor-Smith, J. K., Compas, B. E., Wadsworth, M. E., Thomsen, A. H., & Saltzman, H.
(2000). Responses to stress in adolescence: Measurement of coping and involuntary stress
responses. Journal of Consulting and Clinical Psychology, 68, 976-992.
Deater-Deckard, K. (2001). Annotation: Recent research examining the role of peer relationships
in the development of psychopathology. Journal of Child Psychology and Psychiatry, 42,
Ge, X., Brody, G. H., Conger, R. D., Simons, R. L., & Murry, V. M. (2002). Contextual
amplification of pubertal transition effects on deviant peer affiliation and externalizing
behavior among African American children. Developmental Psychology, 38, 42-54.
Ge, X., Conger, R. D., & Elder, G. H. (2001). Pubertal transition, stressful life events, and the
emergence of gender differences in adolescent depressive symptoms. Developmental
Psychology, 37, 404-417.
Goodyer, I. M., Herbert, J., Tamplin, A., & Altham, P. M. E. (2000). First-episode major
depression in adolescents. British Journal ofPsychiatry, 1 76, 142-149.
Graber, J. A. (2003). Puberty in context. In C. Hayward (Ed.), Gender differences at puberty (pp.
307-321). New York: Cambridge University Press.
Graber, J. A., & Brooks-Gunn, J. (1996). Transitions and turning points: Navigating the passage
from childhood through adolescence. Developmental Psychology, 32, 768-776.
Graber, J. A., Brooks-Gunn, J., & Archibald, A. B. (2005). Links between girls' puberty and
externalizing and internalizing behaviors: Moving from demonstrating effects to
i denti fying p athway s. In D. M. Stoff & E. J. Su sman (Ed s.), Developmental psychobiology
ofuggress.v~ionr New York: Cambridge University Press.
Graber, J. A., Seeley, J. R., Brooks-Gunn, J., & Lewinsohn, P. M. (2004). Is pubertal timing
associated with psychopathology in young adulthood? Journal of the American Academy
of Child and Adolescent Psychiatry, 43, 7 18-726.
Granger, D. A., Weisz, J. R., & Kauneckis, D. (1994). Neuroendocrine reactivity, internalizing
behavior problems, and control-related cognitions in clinic-referred children and
adolescents. Journal ofAbnormal Psychology, 103, 267-276.
Gunnar, M. R. (2000). Early adversity and the development of stress reactivity and regulation. In
C. A. Nelson (Ed.), The effects of adversity on neurobehavioral development: The
M~innesota Symposia on Child Psychology (Vol. 31, pp. 163-200). Mahwah, NJ: Erlbaum.
Gunnar, M. R., & Vazquez, D. M. (2001). Low cortisol and a flattening of expected daytime
rhythm: Potential indices of risk in human development. Development & Psychopathology,
Hampel, P., & Petermann, F. (2006). Perceived stress, coping, and adjustment in adolescents
Journal of Adolescent Health, 38, 409-415.
Hawker, D. S., & Boulton, M. J. (2000). Twenty years' research on peer victimization and
psychosocial adjustment: A meta-analytic review of cross-sectional studies. Journal of
Child Psychology and Psychiatry, 41, 441-455.
Hertsgaard, L., Gunnar, M. R., Larson, M., Brodersen, L., & Lehman, H. (1992). First time
experiences in infancy: When they appear to be pleasant, do they activate the
adrenocortical stress response? Developmental Psychobiology, 25, 319-333.
Hollingshead, A. B. (1975). Four factor index of social status. Unpublished manuscript, Yale
Jaser, S. S., Langrock, A. M., Keller, G., Merchant, M. J., Benson, M., Reeslund, K., et al.
(2005). Coping with the stress of parental depression II: Adolescent and parent reports of
coping and adjustment. Journal ofClinical Child and Adolescent Psychology, 34, 193-205.
Khatri, P., Kupersmidt, J. B., & Patterson, C. (2000). Aggression and peer victimization as
predictors of self-reported behavioral and emotional adjustment. Aggressive Behavior, 26,
Klimes-Dougan, B., Hastings, P. D., Granger, D. A., Usher, B. A., & Zahn-Waxler, C. (2001).
Adrenocortical activity in at-risk and normally developing adolescents: Individual
differences in salivary cortisol basal levels, diurnal variation, and responses to social
challenges. Development & Psychopathology, 13, 695-719.
La Greca, A. M. (2001). Friends or foes? Peer influences on anxiety among children and
adolescents. In W. K. a. T. Silverman, P.D.A. (Ed.), Anxiety disorders in children and
adolescents: Research, assessment and intervention (pp. 159-186): Cambrid ge University
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A
comparison of methods to test mediation and other intervening variable effects.
Psychological M~ethods, 7, 83-104.
Meinlschmid, G., & Hellhammer, D. H. (2003). Two formulas for computation of the area under
the curve represent measures of total hormone concentration versus time-dependent
change. Psychoneuroendocrinology, 28, 9 16-93 1.
Nachmias, M., Gunnar, M. R., Mangelsdorf, S., Parritz, R. H., & Buss, K. (1996). Behavioral
inhibition and stress reactivity: The moderating role of attachment security. Child
Development, 67, 508-522.
Nansel, T. R., Craig, W., Overpeck, M., Saluja, G., Ruan, J., & Health Behaviour in School-aged
Children Bullying Analyses Working Group. (2004). Cross-national consistency in the
relationship between bullying behaviors and psychosocial adjustment. Archives of
Pediatrics and Adolescent M~edicine, 158, 73 0-73 6.
Ohring, R., Graber, J. A., & Brooks-Gunn, J. (2002). Girls' recurrent and concurrent body
dissatisfaction: Correlates and consequences over 8 years. International Journal of~ating
Disorders, 31, 404-415.
Oosterlaan, J., Geurts, H. M., Knol, D. L., & Sergeant, J. A. (2005). Low basal salivary cortisol
is associated with teacher-reported symptoms of conduct disorder. Psychiatry Research,
Petersen, A., & Crockett, L. (1985). Pubertal timing and grade effects on adjustment. Journal of
Youth and Adolescence, 14, 191-206.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects
in simple mediation models. Behavior Research M~ethods, Instruments, and Computers, 36,
Preacher, K. J., & Hayes, A. F. (2006). Asynaptotic and resanpling strategies for assessing and
comparing indirect effects in simple and multiple mediator models. Manuscript submitted
for publication. Pruessner, J. C., Kirschbaum, C.,
Reinhard, C., Wolff, B., & Wadsworth, M. (2006, March). Coping and involuntary stress
response clusters predicting synaptonts of psychopathology for adolescents living in
poverty. Paper presented at the Biennial Meeting of the Society for Research on
Adolescence, San Francisco.
Rovine, M. J., von Eye, A., & Wood, P. (1988). The effect of low covariate criterion correlations
on the analysis-of-covariance. In E. Wegmen (Ed.), Computer science and statistics:
Proceedings of the 20t synaposiunt of the interface (pp. 500-504). Alexandria, VA:
American Statistical Association.
Schmidt, L. A., Fox, N. A., Sternberg, E. M., Gold, P. W., Smith, C. C., & Schulkin, J. (1999).
Adrenocortical reactivity and social competence in seven year-olds. Personality and
Individual Differences, 26, 977-985.
Schwartz, D., McFadyen-Ketchum, S. A., Dodge, K. A., Pettit, G. S., & Bates, J. E. (1998). Peer
group victimization as a predictor of children's behavior problems at home and in school.
Development & Psychopathology, 10, 87-99.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New
procedures and recommendations. Psychological M~ethods, 7, 422-445.
Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural equation models. In S.
Leinhart (Ed.), Sociologicalnzethodology 1982 (pp.290-312). San Francisco: Jossey-Bass.
Southwick, S. M., Vythilingam, M., & Charney, D. S. (2005). The psychobiology of depression
and resilience to stress: Implications for prevention and treatment. Annual Review of
Clinical Psychology, 1, 255-291.
Storch, E. A., Nock, M. K., Masia-Warner, C., & Barlas, M. E. (2003). Peer victimization and
social-psychological adjustment in Hispanic and African-American children. Journal of
Child and Family Studies, 12, 439-452.
Van Bokhoven, I., Van Goozen, S. H. M., Van Engeland, H., Schaal, B., Arseneault, L., Seguin,
J. R., et al. (2005). Salivary cortisol and aggression in a population-based longitudinal
study of adolescent males. Journal ofNeural Transmission, 112, 1083-1096.
Van de Wiel, N., Van Goozen, S. H. M., Matthys, W., Snoek, H., & Van Engeland, H. (2004).
Cortisol and treatment effects in children with disruptive behavior disorders: A preliminary
study. Journal ofAnzerican Academy of Child and Adolescent Psychiatry, 43, 10O1 1 1018.
Wadsworth, M. E., Raviv, T., Compas, B. E., & Connor-Smith, J. K. (2005). Parent and
adolescent responses to poverty-related stress: Tests of mediated and moderated coping
models. Journal of Child and Family Studies, 14, 283-298.
Weichold, K., Silbereisen, R. K., & Schmitt-Rodermund, E. (2003). Short-term and long-term
consequences of early versus late physical maturation in adolescents. In C. Hayward (Ed.),
Gender differences at puberty (pp. 241-276). New York, NY: Cambridge University Press.
Weisberg, H. I. (1979). Statistical adjustments and uncontrolled studies. Psychological Bulletin,
Lisa M. Sontag is a third-year developmental psychology graduate student at the
University of Florida. She received a Bachelor of Science in Psychology at Tulane University.
Her research focuses on adolescent development from a bio-psychosocial orientation. In
particular, her research explores how peer relations, stress reactivity, pubertal timing, and coping
contribute to both healthy adjustment and emotional and behavioral problems. Currently she is
the co-investigator of the APEX (Adolescent Peer Experiences) Study that investigates how peer
relationships influence psychosocial development during the middle school years.