Development and Psychopathology, 7 (1995), 183-192
Copyright © 1995 Cambridge University Press
Printed in the United States of America
Emotion regulation and depressive
symptoms during adolescence:
A functionalist perspective
ROGER KOBAK AND RAYANNE FERENZ-GILLIES
University of Delaware
Abstract
This paper uses a functionalist view of emotion to consider family factors that create risk for depressive
symptoms in adolescents. Two adolescent siblings and their mothers were assessed for emotion regulation
during autonomy-related communication tasks, whereas their mother's attachment- and intimacy-related
functioning was assessed through interview and self-report. The results indicate that older teens' failure to
establish autonomous stances during communication and their mothers' dissatisfaction with intimate
relationships increase risk for teenagers reporting depressive symptoms. Further, mothers' preoccupation with
attachment-related issues was found to be associated with less autonomous communication from older
teenagers. Implications of these findings and future directions for investigating the family factors associated
with adolescent depression are discussed.
*
Depressive symptoms increase dramatically plishment of these tasks. When things go
over the course of adolescence with increas- well, emotions serve valuable motivational
ing prevalence among females (Angold & and communication functions (Kobak,
Rutter, 1992). This change occurs as teenag- Ruckdeschel, & Hazan, 1994). Alterna-
ers face new developmental challenges such tively, when teenagers are unable to pursue
as negotiating autonomy in relationships autonomy-related goals or when their par-
with parents, managing goal conflicts, de- ents have difficulty managing their own
veloping increased intimacy in peer rela- adult concerns, teenagers may become vul-
tionships, and leaving home. Parents and nerable to dysfunctional negative affect and
teenagers must manage the emotions that depressive symptoms. In this paper, we will
accompany these transitions. For their part, use a functionalist view of emotions to in-
parents must manage intimacy in adult rela- vestigate mothers' functioning and their
tionships and update their models of child- teenagers' capacity to regulate emotion as
hood in order to effectively serve as a secure potential risk factors for teenagers' depres-
base for teenage offspring. Teenagers must sive symptoms,
tolerate the uncertainty of asserting their
autonomy-related goals while maintaining .
m t i o n Re u Ia I n n
open lines of communication with parents. E ° 8 i * « pJ i
Emotions may either help or hinder accom- Autonomy-Related Tasks
During the past decade, the study of emo-
tion has gygradually shifted from an
an emphasis
emphasis
was by N MH
L«? ,TT'r J °rant R°3" on
intrapsychic feeling states to a function-
44885 awarded to the first author. ,. .. ., . , . 1Ui"-"u»
allst
Address correspondence and reprint requests to: Perspective that emphasizes the role
Roger Kobak, Department of Psychology, University that emotions play in maintaining orga-
of Delaware, Newark, DE 19716. nized goal-directed behavior. For instance,
183
184 R. Kobak and R. Ferenz-Gillies
Campos emphasized that emotions help to itoring organized goal-directed activity. For
maintain organized activity between the in- instance anger can motivate constructive as-
dividual and the environment on "matters sertion and simultaneously provide parents
of significance to the person" (Campos, with valuable information about their teen-
1994, p. 1). In this perspective, emotions are agers' goals. Thus, emotions serve a valu-
inferred from "1) the way that behavior is able signal function (Kobak et al., in press),
organized, 2) suppositions about what the On the other hand, emotions can easily be-
individual is trying to accomplish, and 3) come dysfunctional in conflictual interac-
determination of whether the striving is tions. For instance, anger may be expressed
progressing smoothly or with difficulty." in accusatory forms that elicit defensive re-
This functionalist view of emotion provides sponses and escalating conflict. To mini-
several useful guidelines for research in de- mize conflict teenagers may disengage from
velopmental psychopathology. First, it interactions and suppress their feelings of
points to using developmental tasks or disappointment and frustration. Subse-
stages as ways of specifying goals that are quently, teenagers' failure to actively pur-
likely to hold particular significance for in- sue attachment and autonomy goals may
dividuals and to organize activity at differ- leave them vulnerable to diffuse negative
ent ages. Developmental goals can be used affect blends and symptomatic or distorted
to identify events that are likely to elicit expressions of emotion (Kobak & Cole,
emotional processes. Second, a functional- 1994).
ist perspective makes it possible to distin- Recent coding systems for examining
guish emotions that organize goal-directed parent-teenager communication distin-
activity from emotions that may interfere guish between functional and dysfunctional
with such activity. By using such criteria, regulation of emotion in similar ways. Ko-
functional and dysfunctional regulation of bak and colleagues have developed a scale
emotion can be defined. Third, dysfunc- for communicative assertiveness that as-
tional regulation of emotion can then be sesses both teenagers' and parents' abilities
used to identify risk factors for psychopath- to effectively communicate about auton-
ology. omy-related goals during discussion of a
Applying this perspective requires that goal conflict (Kobak etal., 1993). They also
researchers identify goals that are likely to developed a secure-base scale that assesses
be significant to a particular subject popu- teenagers' ability to actively communicate
lation. For adolescents, negotiating in- their goals and plans for the future in a
creased autonomy holds particular develop- discussion about their leaving home (Ko-
mental significance (Steinberg, 1990). In bak, Ferenz-Gillies, Everhart, & Seabrook,
examining typical conflicts that occur be- 1994). Allen and colleagues (1990) have de-
tween teens and their parents, Smetana veloped a transcript-based coding system
(1988) noted a marked increase in teenagers' focusing on behavior that promotes auton-
appeal to personal jurisdiction to support omy and relatedness goals during discus-
their positions against parents. Revealed sions of a conflict situation,
differences paradigms have allowed investi-
gators to test teenagers' ability to maintain . .. , ,„ .. „,
e
* . . „.A Attachment and Parenting Teenagers
an autonomous stance during conflict
within the family (Allen & Hauser, Kobak, Parents remain important partners for ado-
Cole, Ferenz-Gillies, & Fleming, 1993). By lescents, yet their caregiving may be in-
examining emotion during these conflicts, fluenced by how successfully they have
investigators can broadly distinguish be- pursued their own attachment and auton-
tween functional and dysfunctional emo- omy-related goals. Recent research suggests
tion regulation. On the one hand, func- that how parents think about their own at-
tional regulation of emotion occurs when tachment history may have important im-
feelings contribute to maintaining and mon- plications for the parent-child relationship
Emotion regulation 185
(VanUzendoorn, Kranenburg, Zwart- that parents' own abilities to maintain au-
Woudstra, Busschbach, & Lambermon, tonomy in the AAI may influence their chil-
1991; VanUzendoorn, 1992). Using the dren's capacity to develop competencies
Adult Attachment Interview (AAI), Main that would protect them from depressive
and Goldwyn (in press) interviewed parents symptoms.
of young children about their own child- A second aspect of parental attachment
hood experiences and analyzed variations in functioning may also have important impli-
interview discourse. Parents were consid- cations for teenagers' adjustment. As teen-
ered secure or free to evaluate attachment agers face the prospect of leaving home,
when their discourse was marked by fluid parents with supportive adult relationships
access to memories and an integrated over- may be more successful in accommodating
view of parent-child relationships. Two teenagers' autonomy needs. These parents
different strategies were evident in the tran- can use their adult attachment relationships
scripts of insecure adults. One group, la- as a secure base that facilitates more flexible
beled dismissing of attachment, showed re- and adaptive parenting. In contrast, if par-
stricted access to attachment information ents are lacking in intimate companionship,
that was reflected in difficulty remembering they are more likely to turn to adolescent
specific interactions or in devaluing their children for support, and a teenager's con-
need to rely on parents. A second group, la- cerns about the parent may create conflict
beled preoccupied with attachment, often with autonomy strivings,
recalled excessively detailed memories of In this article we test a model that views
childhood tinged with anger and had diffi- teenagers' ability to pursue autonomy goals
culty stepping back and integrating these and their mothers'attachment-related func-
experiences into a clear overview of the rela- tioning as risk and protective factors in the
tionship. development of depressive symptomatol-
The relation between parents' AAI strat- ogy. We expect that failure on the teenager's
egies and their capacity to manage emotion part to pursue autonomy-related goals in-
with their offspring has been extensively in- creases risk for depressive symptoms. Teens
vestigated with the parents of infants and who fail to constructively engage in an au-
young children. Crowell and Feldman tonomy task are likely to experience a lack
(1988, 1991) provide evidence linking par- of efficacy. Together, residual anger com-
ents' AAI classifications to parenting be- bined with a perception of failure are likely
havior during a laboratory problem-solving to increase depressive symptoms. We also
task and separation/reunion procedures, anticipate that maternal functioning, in-
They found that secure mothers were gener- eluding attachment strategies in the AAI
ally more supportive in the problem-solving and satisfaction with close relationships,
session, and their children were more coop- will influence a mother's ability to serve as
erative and less negative. Children of preoc- a secure base for her teenager's movement
cupied mothers were generally less compli- toward increased autonomy,
ant and experienced the most negative By studying siblings in the same family
affect during the session. In the separation we can test the developmental specificity of
procedure (Crowell & Feldman, 1991), se- our functionalist framework. That is, we
cure mothers prepared their children most can consider whether our interaction tasks
effectively for separation, although preoc- designed to activate autonomy-related goals
cupied mothers displayed the most anxiety, in adolescents would be more relevant for
In a previous report on the mother-teen- older or younger adolescents. A develop-
aged sibling project, we found that mothers' mental framework suggests that although
preoccupation in the AAI was associated autonomy becomes an important issue be-
with their older teenagers' demonstrating ginning in adolescence, it should become in-
less autonomy in two communication tasks creasingly important as teenagers prepare
(Kobak et al., 1994). These studies suggest to leave home. Thus, we expected that diffi-
186 R. Kobak and R. Ferenz-Gillies
culty asserting autonomy-related goals in which older versus younger adolescents
would be more closely linked with depres- participated was varied according to each
sive symptoms for older teenagers. This ex- family's convenience, however, in 19 fami-
pectation comes from a model of family de- lies the younger teenager attended the first
velopment in which autonomy becomes a session, whereas in 23 families the older
salient task during late adolescence, as teen- teenager attended the first session. Mothers
agers face leaving home (Noller & Calla- and teenagers completed a battery of mea-
han, 1991; Stierlin, 1972) and establishing sures that included interviews, videotaped
greater mutuality with mothers (Youniss & interaction tasks, Q-sorts, and question-
Smollar, 1985). naires. The AAI was administered individu-
ally to each mother.
Methods
Participants Adult A ttachment Interview and Q-sort
Families with two siblings attending the Attachment interviews lasted approxi-
same high school were contacted by mail mately 1 hour and followed the procedure
and asked to participate in a study of par- developed by George, Kaplan, and Main
ent-sibling relationships. Two hundred (1985). Interviews were audiotaped and
sixty-five families received letters, and 54 transcribed. The Attachment Interview Q-
families agreed to participate (20%). Forty- Sort (Kobak et al., 1993), was used to ob-
two mother-adolescent triads were selected tain indexes of mothers' secure/insecure
for the study. The final sample consisted and dismissing/preoccupied tendencies,
mainly of White, middle-class families. The Q-set consists of 100 items that describe
Ninety percent were Caucasian, 8% were features of discourse coherence and work-
Black, and 2% were of Hispanic origin, ing models of attachment figures. After
Model annual income was more than reading a transcript of the interview, two
$40,000 for fathers and was between judges independently sorted the 100 items
$10,000 and $15,000 for mothers. Mothers into nine categories ranging from "most
averaged 14.8 years of education (SD — characteristic" to "least characteristic." Re-
2.2), and fathers averaged 14.7 years (SD liability was determined by computing the
= 3.0). Most households consisted of two- correlation for the two judges using the
parent families. Seventy-four percent of Spearman-Brown formula. In instances
participants were married to their original where reliability was less than 0.58, the
partners, 129b were remarried, and 14% transcript was rated by a third person. The
were unmarried. The number of children in mean composite reliability for the tran-
each family ranged from 2 to 7 with an aver- scripts was 0.77 (range = 0.58-0.90).
age of 2.8. Teenage siblings ranged in age To assess attachment strategies, the at-
from 13 to 19 years, with a mean of 14.7 tachment Q-set items were sorted a priori
years (SD = 0.98) for younger teens and into prototypes (Kobak et al., 1993). A se-
16.9 years (SD = 0.88) for older teens, cure/insecure prototype distinguished be-
Mean age for mothers was 41.5 years (SD tween secure and anxious subjects and a
= 3.8). Gender of teenagers with respect to dismissing/preoccupied prototype distin-
birth order was also similar across families, guished between the two types of insecurity.
Fifty percent of the older teenagers and Mothers' interview ratings were correlated
52% of the younger teenagers were female, with prototype sorts and the resulting cor-
Each family received $60 for participation, relation coefficients indexed secure/inse-
cure and dismissing/preoccupied strategies.
Procedure These two dimensions have been found to
This study was part of a larger data collec- discriminate between Main and Goldwyn's
tion project. Mothers attended two ses- three major attachment categories, secure,
sions, one with each adolescent. The order dismissing, and preoccupied (Kobak et al.,
Emotion regulation 187
1993). Correlations with the AAI security/ Communicative Assertiveness. Subjects
insecurity prototype ranged from - 0.81 to who communicated high levels of contempt
0.75 with a mean of - 0.04 (SD = 0.43) or frustration through nonverbal behavior
and the dismissing/preoccupied prototype such as sighing and withdrawal and more
ranged from - 0 . 5 7 to 0.71 with a mean of overt attacking behaviors such as raising
- 0 . 0 5 (SD = 0.35). These values are simi- voice level and criticizing the partner re-
lar to those obtained by Kobak et al. (in ceived high scores on the Dysfunctional
press) with an adolescent sample. Anger Scale. Interrater reliability (Pearson
r correlation coefficient) for the Dysfunc-
Future discussion task and ratings tion A n 8 f Scale was .73. High scores for
Communicative Assertiveness were given to
Mothers participated with each adolescent subjects who effectively communicated
in a ten-minute videotaped future discus- t h e ir goals and point of view about the dis-
sion task. Teenagers and their mothers were agreement. Low scores were assigned to
given the following instructions: family members who talked little or who
agreed with the partner without presenting
Imagine that (teen's name) is ready to leave t h d r p o i n t o f y i e w > I n t e r r a t e r reliability
home, discuss (teen's name)'s goals and plans for ( P e a r s o n r c o r r e l a t i o n coefficient) for the
the future and how your relationship with each D functional A Scale was ^
other is going to change.
Secure base scale. Raters evaluated the Mothers'reports of satisfaction with
teenagers' interactions using the Secure close relationships
Base Scale. This scale assesses three aspects
of conversation that reflect the teens secu- Mothers completed several self-report mea-
sure
rity or confidence in self: emotional engage- * °* their satisfaction in close relation-
ment, confidence in self and goals, and con- ships. Perceived competence in social, inti-
fidence in the relationship (Kobak et al., mate and nurturing relationships subscales
fr
1994). The 7-point scale combines nonver- °m the Self-Perception Profile for Adults
bal and verbal cues. Low ratings were as- (Harter, 1988) were converted to z scores
signed to teenagers who showed minimal and composited with z scores on the Univer-
slt
engagement in the task and failed to discuss V o f California-Los Angeles (UCLA)
plans for the future. Midpoint was assigned loneliness scale (Russell, Peplau, & Cu-
trona
to teenagers who showed some engagement » 1980)-
in the task, but failed to discuss both future
plans and relationship changes. High rat- Teenagers'self-reports of
ings were assigned to teens who displayed depressive symptoms
active involvement and balanced discus-
sions of future goals with sensitivity to rela- Teenager's completed the Dimensions of De-
tionship changes. Raters, blind to other pression Profile for Adolescents (DDPA;
measures in the study, independently as- Harter, Marold, & Nowakowski, 1987).
T h i s is a
signed a score to the adolescent's behavior 30-item scale designed to assess five
for each third of the 10-min discussion. In- theoretically derived dimensions of depres-
sion in
terrater reliability (Pearson r correlation co- children and adolescents: (a) Dys-
efficient) for the Secure Base Scale was .75. P h o r i c M o o d . ( b ) Depressed Energy/Inter-
est, (c) Global Self-Worth, (d) Self-Blame,
_ . . ,. , .. , . and (e) Suicidal Ideation. The questionnaire
Ratings of emotion regulation during , a structured aUernative format de.
mother-teenage conflict s i g n e d t 0 o v e r c o m e t h e subject>s tendency
Independent teams of raters coded the teens to give socially desirable responses. The
behavior during problem-solving on one of DDPA has been extensively tested with ele-
two scales: a Dysfunctional Anger and mentary and secondary school students.
188 R. Kobak and R. Ferenz-Gillies
Subscale internal reliability ranges from tion in the problem-solving and future
0.77 for self-blame to 0.90 for suicidal ide- discussion tasks. Ratings of mothers' dis-
ation. Correlations between subscales range missing/preoccupation covaried with their
from 0.30 for self-blame and energy to 0.82 oldest teenager's communicative assertive-
for mood and self-worth. Self-reports on ness during problem-solving and showed a
the DDPA have been concurrently vali- similar trend for youngest teenager's corn-
dated with teachers' ratings (Harter et al., municative assertiveness. These correla-
1987). tions suggest that mothers' who were more
preoccupied in the AAI had teenagers who
R .. were less able to assert their goals during the
problem discussion. A similar association
The mean depression scores for older ado- was evident for mothers' dismissing/preoc-
lescents (x = 2.9) and younger adolescents cupation and their oldest teenagers' discus-
(x = 3.0) were not significantly different, sion of goals and plans for leaving home in
Depression scores are based on the mean the future discussion task. Thus, mothers
item placement on the Dimensions of De- who were preoccupied in the AAI had old-
pression Profile. Subjects' responses to est teenagers who were less autonomous
items range from 1 indicating that descrip- during problem-solving and discussions of
tions of depressed mood, self-blame, low leaving home. However, neither mothers'
energy, and low self-worth are really true security in the AAI nor their satisfaction
for me to 4 indicating the absence of de- with close relationships showed significant
pressed symptoms are really true for me. associations with teenagers' emotion regula-
Overall, both younger and older teens tion.
showed normative levels of depressive
symptoms compared with the inpatient ado- ._, , ... - .. .
, . ... f. . , ,. r J Maternal personality functioning,
lescents with clinical diagnoses of depres- „. ... , . .
. . , , . j , , . emotion regulation, and depressive
sion who reported elevated levels off depres- . . , , :'
/ - = 2.3).
sion (x X ?\ These
i-u • 4.- *
investigators report* symptoms in oldest teenagers
that a large sample of middle-school stu- Regression analysis was used to test how
dents had a mean score of 3.1 onthisinven- oldest teenagers' emotion regulation and
tory indicating lower levels of depressive mothers' personality functioning were re-
symptom compared with the clinical group lated to depressive symptoms among the
(Renouf & Harter, 1990). older siblings in each family. To test the in-
fluence of maternal personality and teenag-
~, .,.!., , i ers' emotion regulation on teenagers' de-
Covanation between maternal . . " * i J*
AAI strategies and teenaeers' pressive symptoms, maternal and teenage
. .. variables were initially entered separately,
emotion regulation .. , . .. , . > *
then entered simultaneously to evaluate cu-
Because both mother variables and teenag- mulative effects. Oldest teenager's depres-
ers' emotion regulation were to be used as sive symptoms were regressed on maternal
independent variables in accounting for personality variables in the first column of
teenagers' depressive symptoms, the corre- Table 2. The overall regression model was
lations between these variables were exam- significant with mothers' lack of satisfac-
ined in Table 1. Previous studies using the tion with close relationships and mother's
AAI have reported associations between use of preoccupied as opposed to dismissing
maternal AAI strategies and child behavior strategies increasing teenagers' risk for de-
in problem-solving situations (Crowell & pressive symptoms. In the second column
Feldman, 1988). Examination of the corre- of Table 2, oldest teenager's depressive
lations indicated that only the AAI dismiss- symptoms were regressed on teenager's
ing/preoccupied dimension was correlated emotion regulation during problem solving
with measures of teenagers' emotion regula- and future discussion tasks. The overall re-
Emotion regulation 189
Table 1. Correlations between maternal variables and teenager's
behavior during communication tasks
Mothers' Personality
AAI AAI Satisfaction
(Security/ (Dismissing/ With Close
Anxiety) Preoccupied) Relationships
Oldest teenager's behavior
Communicative assertiveness -0.14 0.39* 0.05
Anger 0.07 0.16 0.11
Secure base 0.01 0.46** 0.05
Youngest teenager's behavior
Communicative assertiveness 0.11 0.23*** -0.18
Anger -0.05 0.00 -0.18
Secure base 0.01 0.14 0.01
*p < .01.
**p < .001.
***p < .10.
Table 2. Oldest teenager's depressive symptoms regressed on maternal personality and
oldest teenager's communication
Oldest Teenager's Oldest Teenager's Oldest Teenager's
Depressive Depressive Depressive
Symptoms Symptoms Symptoms
B r B r B
Gender 0.13 .09 0.27 0.31*
Maternal personality
AAI security -0.19 -.21* -0.21
AAI rep./preoccup. -0.33** - . 3 1 * * -0.12
Relationship satisfaction -0.32** -.39*** -0.31**
Teenager communication
Communicative assertiveness -0.40** -.40*** -0.43**
Dysfunctional anger -0.13 -.09 -0.07
Secure base -0.18 -.26** -0.08
Multiple/? = 0.50, Multiple R = 0.50, Multiple/? = 0.64,
F = 3.02** F = 2.95** F = 3.28***
*p < .10.
**p < .05.
***p < .01.
gression model was significant with teenag- were regressed on depressive symptoms in
er's communicative assertiveness during column 3 on Table 2. The overall model re-
problem solving reducing the likelihood of mained significant with both teenagers' lack
depressive symptoms. Finally, to test the of communicative assertiveness and their
relative contributions of maternal personal- mothers' lack of satisfaction with intimate
ity and teenage emotion regulation both the relationships contributing significantly to
maternal and teenage blocks of variables depressive symptoms.
190 R. KobakandR. Ferenz-Gillies
Table 3. Youngest teenager's depressive symptoms regressed on maternal
personality and youngest teenager's communication
Youngest Youngest Youngest
Teenager's Teenager's Teenager's
Depressive Depressive Depressive
Symptoms Symptoms Symptoms
B r B r B
Gender 0.05 .11 0.27 0.08
Maternal personality
AAI security -0.01 -.01 0.00
AAI rep./preoccup. -0.00 -.06 -0.03
Relationship satisfaction 0.23 .19 0.21
Teenager communication
Communicative assertiveness -0.34* -.24** -0.29
Dysfunctional anger 0.25 .23** 0.28
Secure base 0.05 - .05 0.02
Multiple/? = 0.24, Multiple/? = 0.38, Multiple/? = 0.42,
F = 0.71 F = 1.51 F = 1.03
*p < .05.
**p < .10.
Maternal personality functioning, their parents. The functionalist approach
emotion regulation and depressive pointed us toward identifying personally
symptoms in youngest teenagers significant goals for adolescents and their
mothers. By specifying these autonomy-
A similar regression model predicting related goals for teenagers and attachment-
younger teenagers' depressive symptoms is and intimacy-related goals for their moth-
presented in Table 3. Maternal variable re- ers, we were guided toward developing
ported in column 1 did not account for theoretically relevant assessments that in-
significant variance in younger teenagers' cluded parent-teenager communication
depressive symptoms. Similarly, teenage tasks (Kobak & Duemmler, 1994) as well as
emotion regulation variables reported in the interview and self-report measures of moth-
second column did not produce an overall ers' functioning. Further, our functionalist
significant effect. However, examination of framework suggested that emotion serves
the individual beta weights in column 2 does an important monitoring functioning as in-
indicate that the contribution of younger dividuals pursue their goals. Thus, frus-
teenagers' communicative assertiveness was trated pursuit of these goals would create
significant. Not surprisingly, when the potential risk for negative affect and de-
younger teenagers' depressive symptoms pressive symptoms in adolescents. Our re-
was regressed on both maternal personality sults support this hypothesis most clearly in
and teenage emotion regulation variables in the case of older teenagers. More specifi-
column 3, neither the overall regression cally, older teenagers who had difficulty as-
model nor any of the individual beta serting their own viewpoint or autonomy
weights accounted for significant variance. during a conflict discussion and whose
mothers were dissatisfied with their own in-
timate adult relationships reported in-
Discussion creased levels of depressive symptoms.
The findings indicate that a functionalist Mothers' satisfaction with her close adult
approach provides a valuable framework relationships was consistently associated
for studying emotion in adolescents and with reduced reports of depressive symp-
Emotion regulation
toms by older teenagers. This finding is con- tial effects on teenagers' emotion regulation
sistent with the Bowlby's (1988) notion that during conversation. Kobak and colleagues
close relationships continue to serve an im- (1994) have suggested that preoccupied
portant attachment function across the life- mothers' may find their oldest teenagers
span. At a speculative level it appears that moves toward autonomy anxiety provoking
when mothers feel support in their adult re- and, consequently, may redirect discussion
lationships, they can function more ade- involving the teenager's autonomy toward
quately as a secure base as their teenaged maternal concerns and worries about the
offspring prepare to leave home. Alterna- teenager leaving home. Growell and Feld-
tively, when mothers are dissatisfied with man (1988) have reported similar effects of
their close relationships, a greater burden maternal preoccupation in the AAI on re-
may be placed on older children who may ducing toddlers' problem-solving behavior,
become alternative sources of support for By reducing teenagers' ability to establish
their parents (Zahn-Waxier, Cole, & Bar- an autonomous stance in the family, these
rett, 1991). Older teenagers in this situation parents may increase the teenagers' negative
may be faced with stresses associated with affect and risk for depressive symptoms
the normal challenge of leaving home as (Kobak & Cole, 1994).
well as by concerns about their mothers' Although the findings in this study sup-
abilities to function without them. port a functionalist view of emotion regula-
For the most part, the pattern of rela- tion and its role as a risk factor in adoles-
tionships between variables was less robust cent depression, future studies need to
for younger than for older teenagers. In replicate and further assess the family con-
part, this result is consistent with our frame- text within which depression develops. For
work that suggest that teenagers' auton- instance, the effects of emotion regulation
omy-related goals may hold particular emo- on older teenagers' depressive symptoms
tional significance as they approach late are open to varying interpretations. On the
adolescence and the prospect of leaving one hand, these effects may be the product
home. The difference between older and of the age of the older adolescent and the
younger teenagers within the same family salience of the communication tasks that we
highlights how assessment of emotion regu- used in our study. On the other hand, our
lation must rely on tasks that have particu- findings may be attributable to birth order
lar developmental significance for the indi- effects. It may be that both maternal and
viduals involved. Despite the relative lack emotion-regulation variables are effected
of findings for younger teenagers, one vari- by higher levels of anxiety that often char-
able, communicative assertiveness, did acterize parent-child interactions of first-
show similar relations for both siblings, born offspring. A larger sample would be
Thus, when teens could actively assert their required to test these alternative interpreta-
autonomy in a conflict discussion, they tions.
were less likely to report depressive symp- A study using a larger sample of adoles-
toms cents could also explore how gender moder-
Mothers' AAI strategies appeared to ates the relation between maternal and emo-
have relatively little direct effect on either tion regulation variables. We anticipate that
siblings' depressive symptoms. However, the kinds of processes indentified with our
our correlational analysis indicated that functionalist framework might hold partic-
mothers with preoccupied strategies in the ular promise for accounting for the devel-
AAI were likely to have oldest teenagers opment of depressive symptoms in female
who demonstrated less autonomy in both adolescents. Several studies support this hy-
the conflict and the future discussion tasks, pothesis. First, daughters may have more
Thus, it is likely that mothers' preoccupa- difficulty than sons in struggling with em-
tion in the AAI is indirectly linked to teen- pathy and guilt in relation with mothers
agers' depressive symptoms through poten- who have difficulty supporting their auton-
192 R. KobakandR. Ferenz-Gillies
omy (Zahn-Waxier, Cole, & Barrett, 1991). (Youniss & Smollar, 1985) and, conse-
Second, adolescent daughters engage their quently, the difficulties with their mothers
mothers more frequently and with more in- may have more deleterious consequences
timate disclosure than do adolescent sons than with sons.
References
Allen, J., & Hauser, S. (1992). Autonomy and related- tionships (Vol. 5, pp. 121-149). London: Jessica
ness in adolescent-family interactions as predictors Kingsley.
of qualities of models of attachment in young Kobak, R., Ruckdeschel, K., & Hazan, C. (1994).
adulthood. Unpublished manuscript, University of From symptom to signal: An attachment view of
Virginia, Department of Psychology. emotion in marital therapy. In S. Johnson & L.
Allen, J., Hauser, S. I., O'Connor, T. G., Bell, K. L., Greenberg (Eds.), Emotions in marital therapy.
& Eickholt, C. (1990). The autonomy relatedness New York: Brunner/Mazel.
coding system: A scoring manual. Unpublished Main, M., & Goldwyn, R. (in press). Adult attachment
manuscript, University of Virginia, Charlottesville. classification system. Unpublished manuscript,
Angold, A., & Rutter, M. (1992). Effects of age and pu- University of California, Berkeley, Department of
bertal status on depression in a large clinical sample. Psychology.
Development and Psychopathology, 4, 5-28. Noller, P., & Callahan, C. (1991). The adolescent in
Bowlby, J. (1988). The secure base. New York: Basic the family. London: Routledge.
Books. Renouf, A. G., & Harter, S. (1990). Low self-worth
Campos, J. (1994, Spring). The new functionalism in and anger as components of the depressive experi-
emotion. SRCD Newsletter. ence in young adolescents. Development and Psy-
Crowell, J., & Feldman, S. (1988). Mothers* internal chopathology, 2, 293-310.
models of relationships and children's behavioral Russell, D., Peplau, L., & Cutrona, C , (1980). The re-
and developmental status: A study of mother-child vised UCLA Loneliness Scale: Concurrent and dis-
interactions. Child Development, 59, 1273-1285. criminant validity evidence. Journal of Personality
Crowell, J., & Feldman, S. (1991). Mothers' working and Social Psychology, 39, 472-480.
models of attachment relationships and mother Smetana, J. (1988). Concepts of self and social con-
and child behavior during separation and reunion. vention: Adolescents' and parents' reasoning about
Developmental Psychology, 27, 597-605. hypothetical and actual family conflicts. In M.
George, C , Kaplan, N., & Main, M. (1985). The at- Gunnar & W. A. Collins (Eds.), Minnesota sympo-
tachment interview for adults. Unpublished manu- sium on child psychology (Vol. 21, pp. 79-122).
script, University of California, Berkeley, Depart- Hillsdale, NJ: Erlbaum.
ment of Psychology. Steinberg, L. (1990). Autonomy, conflict and har-
Harter, S. (1988). Manualfor the Self-Perception Pro- mony in the family relationship. In S. S. Feldman &
file for Adults. University of Denver, Department G. Elliott (Eds.), At the threshold: The developing
of Psychology. adolescent (pp. 255-276). Cambridge: Harvard
Harter, S., Marold, p . , & Nowakowski, M. (1987). University Press.
Manual for the dimensions of depression profile Stierlin, H. (1972). Separating parents and adoles-
for children and adolescents. University of Delfiver, cents. New York: Aronson.
Department of Psychology. VanUzendoorn, M., Kranenburg, M., Zwart-Wouds-
Kobak, R., Cole, H., Ferenz-Gillies, R., & Fleming, tra, A., Busschbach, A., & Lambermon, W.
W. (1993). Attachment and emotion regulation (1991). Parental attachment and children's socio-
during mother-teen problem-solving: A control emotional development: Some findings on the va-
theory analysis. Child Development, 64, 231-245. lidity of the adult attachment interview in the Neth-
Kobak, R., & Cole, C. (1994). Attachment and meta- erlands. International Journal of Behavioral
monitoring: Implications for adolescent auton- Development, 14, 375-394.
omy and psychopathology. In D. Cicchetti (Ed.), VanUzendoorn, M. H. (1992). Intergenerational
Rochester symposium on development and psycho- transmission of parenting: A review of studies in
pathology Vol. 5: Disorders of the self (pp. 267- nonclinical populations. Developmental Review,
297). Rochester, NY: Rochester University Press. 12, 76-99.
Kobak, R., Ferenz-Gillies, R., & Everhart, E., Sea- Youniss, J., & Smollar, J. (1985). Adolescents rela-
brook, L. (1994). Maternal attachment strategies tions with mothers, fathers, and friends. Chicago:
and emotion regulation with adolescent offspring. University of Chicago Press.
Journal of Research on Adolescence, 4, 553-566. Zahn-Waxler, C , Cole, P. M., & Barrett, K. C.
Kobak, R., & Duemmler, S. (1994). Attachment and (1991). Guilt and empathy: Sex differences and im-
conversation: Toward a discourse analysis of ado- plications for the development of depression. In J.
lescent and adult security. In D. Perlman & K. Garber & K. A. Dodge (Ed.), The development of
Bartholomew (Eds.), Attachment processes in emotion regulation and dysregulation (pp. 243-
adulthood: Advances in the study ofpersonal rela- 272). Cambridge: Cambridge University Press.
