Comorbidity and longitudinal associations among anxiety, depression and aggression in adolescence: A

1
Comorbidity and longitudinal associations among
anxiety, depression and aggression in
adolescence Anxiety as a mediating processEric
Buhs, Ph.D., Susan M. Swearer, Ph.D., Amanda B.
Siebecker, M.A., Kelly Brey Love, M.A.University
of Nebraska-Lincoln
RESULTS Path analyses (MPLUS regressions, see
Figure 2) indicated that the model linking
Bully/Victim status and Aggression (increase) fit
the data well (RMSEA.03, CFI.99) and indicated
that Anxiety mediated this relationship. Results
indicated that Depression and Anxiety were
linked, but Depression did not predict Aggression
independently (i.e. when controlling for
Anxiety). Additionally, Victim status was the
only relational category that significantly
predicted Anxiety. A model linking
Bully/Victim status to Aggression via Depression
(i.e. Depression as mediator) could not be fitted
to the data.

• INTRODUCTION
• Current research regarding bullying and
  victimization often suggests that these
  relationships and interactions are stressful
  relational contexts for children.
• In the literature there is a seeming paradox in
  that research findings suggest that experiences
  of both bullies and victims may play a causal
  role in increased internalizing disorders
  (Kumpulainen et al., 1998).
• Further complicating matters, research on
  internalizing disorders has established that
  depressive and anxious symptoms are often
  comorbid and may be associated with higher levels
  of aggression or externalizing problems (Quiggle
  et al., 1992 Zahn-Waxler et al., 2000). This
  suggests, sadly, that even victimized children,
  as well as bullies, are at risk for internalizing
  problems and increased aggression.
• Conceptual Frameworks
• Extensive longitudinal work on the developmental
  pathways for bully/victim interactions and
  associated internalizing/externalizing problems
  is still lacking. However, broader conceptual
  frameworks that present a more general set of
  premises may be empirically evaluated to
  establish whether or not these causal linkages
  exist.
• Within life-stress or relational stress models
  of psychopathology (Albee, 1984 Coie, 1990
  Rudolph et al., 2000), stressful relationships
  are expected to increase the frequency and/or
  severity of pathology. Prior findings have
  indicated both bullies and victims experience
  greater depressive and anxious symptomatology.
• Studies of internalizing disorders as predictors
  of increased aggression (Brendgen et al., 2002
  Renouf Harter, 1990) also suggest that as
  abusive relational contexts yield negative social
  experiences, children develop negative
  expectations for future social interactions and
  attribute hostile intent to peers (Zahn-Waxler et
  al., 2000).
• While children with externalizing styles appear
  more likely to act aggressively based on hostile
  attributions, we hypothesize that children with
  internalizing behavioral styles are also likely
  to display increased aggression. This contrasts
  with models suggesting non-overlapping types of
  social responses to peer abuse for bullies vs.
  victims.
• The Current Model

Figure 1 Conceptual Model
Relational
Increased
Internalizing
Stress
Externalizing
problems
Problems
Depressive
Aggressive
Bully/victim
Anxious
interactions
Behavior
Symptoms

• DISCUSSION
• While depression and anxiety symptoms were
  linked (comorbid) in this sample, only anxiety
  mediated the link between Bully/Victim status and
  increased aggressive behaviors. This suggests
  that, within the current data set, children who
  are victimized tended to display greater
  internalizing problems, but only higher levels of
  anxiety are predictive of relative increases in
  aggression.
•  Cyclical increases may occur in maladaptive
  adjustment and behavior patterns within social
  environments that support or tolerate BV
  interactions, especially when larger proportions
  of adolescents present are victimized.
• Current findings are limited by reliance on
  self-report data and a sample N that precluded
  standard SEM analyses or more detailed
  examinations of growth trajectories in
  Aggression.

• METHOD
• Active parental consent and youth assents were
  obtained for each participant in the study.
  Participants completed the instruments each
  spring across three years of middle school entry
  year, mid year, and exit year.
• Participants
• In 2000-2003 data were collected with 138 middle
  school students from a larger longitudinal study
  on bullying and victimization. 54 of
  participants were females (n 75 males n 63).
  Racial distribution was 79 Caucasian (n
  110), 5 African American (n 7), 3.5
  Latino/Hispanic (n 5), 4.3 Asian/Asian
  American (n 6), 2.1 Native-American (n 3),
  and 5 mixed minority (n 7).
• Bully/victim status distribution at entry was
  victims (n 53) bullies (n 9) bully-victims
  (n 31) not involved (n 17) bystanders (n
  41)
• Measures
• Childrens Depression Inventory (CDI) 27-item
  self-report of depression. Items assess overt
  symptoms of depression using a 3-point scale ( 0
  - 2). The CDI has demonstrated acceptable
  internal consistency, reliability and convergent
  validity (Kovacs, 1992). Alphas ranged from .85
  to .87 across the three years.
• Bully Survey-Student Version (BYS-S, Swearer,
  2001) asks students about bullying, perceptions
  of bullying, and attitudes toward bullying.
  Bullying is defined as teasing, saying mean
  things, or leaving someone out of a group to
  physical attacks (hitting, pushing, kicking)
  where one person or a group of people picks on
  another person over a long time. Bullying refers
  to things that happen in school but can also
  include things that happen on the school grounds
  or going to and from school. Also includes
  questions about victimization, observations of
  bullying (bystander), and about bullying others
  (time frame is within the last year).
• Multidimensional Anxiety Scale for Children
  (MASC) 39-item self-report that assessing
  anxiety in children. A 5-item subset of the
  physical symptoms subscale was used to indicate
  anxious symptomatology. Alpha was .77 (T2).
• Aggression Questionnaire a 34-item self-report
  measure that assesses anger and aggression. A
  5-item subset of the AQ physical aggression scale
  was used to indicate levels of aggressive
  beahvior. Higher scores on this measure indicated
  increased aggression, relative to predicted
  values, from Time 1 to Time 3. Alphas ranged
  from.87 to .85 across the three years.

Figure 2 Results
Relational
Increased
Internalizing
Stress
Externalizing
problems
Problems
Aggressive
Anxious
Behavior
Bully
Symptoms
.31
(change)

.43
Victim
.54
ns
Bully-Victim
Depressive
Symptoms
Bystander