Resilience as Process

Abstract

Resilience, the development of competence despite severe or pervasive adversity, is examined using data from a longitudinal study of high-risk children and families. The study is guided by an organizationaldevelopmental perspective. Resilience is conceived not as a childhood given, but as a capacity that develops over time in the context of person-environment interactions. Factors related to resilience in our study are examined in terms of this transactional process. From our studies, we have found emotionally responsive caregiving to mediate the effects of high-risk environments and to promote positive change for children who have experienced poverty, family stress, and maltreatment. The implications of these findings are discussed.

Full text

Development
and
Psychopathology,
5 (1993), 517-528
Copyright © 1993 Cambridge University Press
Printed in the United States of America
Resilience as process
BYRON EGELAND, ELIZABETH CARLSON,
AND
L. ALAN SROUFE
Institute of
Child
Development,
University
of Minnesota
Abstract
Resilience, the development of competence despite severe or pervasive adversity, is examined using data from
a longitudinal study of high-risk children and families. The study is guided by an organizational-
developmental perspective. Resilience is conceived not as a childhood given, but as a capacity that develops
over time in the context of person-environment interactions. Factors related to resilience in our study are
examined in terms of this transactional process. From our studies, we have found emotionally responsive
caregiving to mediate the effects of high-risk environments and to promote positive change for children who
have experienced poverty, family stress, and maltreatment. The implications of these findings are discussed.
The most recent wave of risk research
proj-
ects has focused on resilience and attempted
to identify protective factors and processes
associated with positive developmental out-
comes (Werner,
1990).
Despite variability in
the definition of resilience across studies,
we have learned much about factors that
mediate the relation between adversity and
more positive child adaptation. We have
moved away from more linear models to a
better understanding of the complexities in-
volved in the relations among risk and pro-
tective factors and the developing child and
the multiple pathways to both adaptive and
maladaptive outcomes (Cicchetti, Toth, &
Lynch, in press).
Resilience has been described as the ca-
pacity for successful adaptation, positive
functioning, or competence (Garmezy,
1993;
Masten, Best, & Garmezy, 1990) de-
spite high-risk status, chronic stress, or fol-
The Mother-Child Project is currently supported by a
grant from the National Institute of Mental Health
(MH40864).
We
also wish to thank the Irving
B.
Harris
Foundation for their support.
Address correspondence and reprint requests to:
Byron Egeland, Institute of Child Development, Uni-
versity of Minnesota,
51
East River Road, Minneapo-
lis,
MN 55455.
lowing prolonged or severe trauma. Resil-
ience is often operationalized as the positive
end of the distribution of developmental
outcomes in a sample of high-risk individu-
als (Rutter, 1990). While these definitions
generally are accepted by researchers of risk
and resilience, factors defining risk samples
and definitions of adaptation and compe-
tence vary widely across studies. In part,
this variability may be resolved by a shift in
research emphasis: from the study of risk
variables to that of risk and protective
mechanisms and the process of "negotiating
risk situations" (Rutter, 1990).
Our own research has increasingly led us
to view resilience in terms of a transactional
process within an organizational frame-
work. From this perspective, develop-
mental outcomes are determined by the
interaction of genetic, biological, psycho-
logical, and sociological factors in the con-
text of environmental support (Cicchetti &
Schneider-Rosen, 1986; Sroufe, 1979). Ac-
cording to this view, any constitutional or
environmental factors may serve as vulnera-
bility, protective, or risk variables, directly
or indirectly influencing behavior. The de-
velopmental process is characterized by a
hierarchical integration of behavioral sys-
tems whereby earlier structures are incorpo-

518
B.
Egelandetal.
rated into later structures in increasingly
complex forms. The individual actively par-
ticipates in this process, bringing to new ex-
perience attitudes, expectations, and feel-
ings derived from a history of interactions
that, in turn, influence the manner in which
environmental cues and stimuli are interpre-
ted and organized. In keeping with this
view, early experience is of critical impor-
tance in shaping the way later experience is
organized. The organizational perspective
provides a useful model for conceptualizing
the complex relationships among risk and
protective mechanisms, prior adaptation,
and resilience.
Within an organizational-developmental
framework, resilience or competence is
viewed as the ability to use internal and
external resources successfully to resolve
stage-salient developmental issues (Waters
& Sroufe, 1983). Competence in resolving
issues in one developmental period does not
predict later competence in a linear deter-
ministic way; rather, competence at one pe-
riod is thought to make the individual
broadly adapted to the environment and
prepared for competence in the next period
(Sroufe & Rutter, 1984). The way in which
early developmental tasks are resolved are
thought to serve a strong and enduring risk
or protective function. Developmentally
relevant issues in the early years include the
formation of an effective attachment rela-
tionship (the 1st year of life) and effective
autonomous functioning (age 2). Major is-
sues during the preschool period include an
expanded ability to organize and coordinate
environmental resources, an interest in en-
gaging problems posed by the environment,
and effective peer relations and emotional
self-regulation. The formation of loyal
friendships and effective peer group func-
tioning are salient issues in middle child-
hood (Cicchetti & Schneider-Rosen, 1986;
Sroufe, 1989).
In our research, individuals at risk for
developmental problems due to adversity
and individuals who forge more successful
adaptations despite troubled beginnings of-
ten have been found to have environments
that are supportive in critical ways or that
change in positive ways prior to the child's
recovery. We have found that some individ-
uals recover more readily following periods
of maladaption and/or are more resistant
to the negative impacts of environmental
stress (Farber & Egeland, 1987); however,
even here, we find such individuals best de-
scribed by history of adaptation over time
and across phases of development. Rather
than being a childhood given or a function
of particular traits, the capacity for resil-
ience develops over time in the context of
environmental support.
In the next section, we will describe our
longitudinal project and our findings on
risk, resilience, and protective mechanisms.
We will discuss these findings from an orga-
nizational perspective and in relation to re-
sults from other longitudinal research on re-
silience.
Mother-Child Project: A Longitudinal
Study of High-Risk Children
Since 1975, we have conducted a longitudi-
nal study of high-risk children and their
families. A sample of
267
women in the last
trimester of their first pregnancies and who
fell below the poverty line were recruited
through local public health clinics to partic-
ipate in the study. Poverty was the major
risk factor for these families, although there
were many associated risk conditions in-
cluding mother's age at the time of birth
(mean age = 20.5 years, range 12-37 years),
marital status (62% were single at the time
of the birth of their child, and only 13% of
the biological fathers were in the house
by the time the child was 18 months), and
education (40% had not completed high
school). Many were multiproblem families
living in chaotic and disruptive environ-
ments. For example, the incidence of child
abuse and neglect in this sample was more
than 15%, and drug and alcohol abuse was
high.
Over the past
18
years, we have collected
detailed and comprehensive measures of
child adaptation at each developmental pe-
riod starting in infancy and continuing
through age 18. The data support the risk

Resilience as process
519
status of the sample. Eighty percent of the
children received some form of special edu-
cation services in elementary school, and
18%
were retained during this period (Ege-
land & Abery, 1991). Approximately 20%
of the children in the first grade fell in the
clinical range (score >65) on teacher rat-
ings on the Internalizing or Externalizing
Scale of the Child Behavior Checklist. De-
spite the large proportion of poorly func-
tioning children, at each assessment we
found some children who were functioning
well above what would be predicted based
on the degree of risk they experienced.
In keeping with an organizational per-
spective, emphasis in our research has been
placed on identifying and examining mean-
ingful patterns of behavior rather than
specific outcomes. Continuity over time is
expected to be found at the level of organi-
zation of behavior rather than in identical
behaviors across assessments (Sroufe,
1979).
Such an approach requires the use of
broadband measures that tax the capacity
of the child to coordinate social, emotional,
and cognitive systems in different situations
(Waters
&
Sroufe, 1983).
Confidence in such measures depends on
assessments involving multiple situations
and procedures and, when possible, multi-
ple sources of information. In our study,
mothers provided information about the
environment in which the children develop;
however, they were not used as sources of
information about children's adaptation.
Instead, we obtained observations in the
home and in structured laboratory situa-
tions,
school assessments including teacher
ratings and independent school observa-
tions by members of our
staff,
and child in-
terviews and testing.
In addition to information on child ad-
aptation, we obtained information on pa-
rental characteristics (e.g., IQ, depressive
symptomology, parental beliefs), child
characteristics (e.g., temperament during
infancy, IQ at later ages), mother-child in-
teraction; family life circumstances (e.g.,
stressful life events, family support, rela-
tionship status), and quality of the home en-
vironment (e.g., HOME [Caldwell & Brad-
ley, 1984]). These data were used to account
for child adaptation or as moderator vari-
ables,
to account for deflections in develop-
mental adaptation. The assessments began
during the last trimester of the mother's
pregnancy and continued up to the current
assessment at age 18. During the first year
of the child's life, assessments of tempera-
ment occurred in the newborn nursery and
at Days 5 and 10. Interviews, testing, and
observation of mother-infant interaction
were conducted at 3, 6 (twice), 9, and 12
(twice) months. Assessments were con-
ducted at 6-month intervals during the tod-
dler and preschool period and at the end of
kindergarten and first, second, third, and
sixth grades in the elementary school years.
A family observation was conducted at age
13,
and individual assessments were admin-
istered at ages 16,17
!£,
and 18. In addition,
a subsample participated in a preschool and
was observed everyday; the same subsample
was observed in a summer camp at approxi-
mately age 11 and at a camp reunion at ap-
proximately age 15.
The Impact of Poverty
In general in our study, we found poverty
and factors associated with poverty to have
had a pervasively negative effect on child
adaptation. We found many children living
in poverty to be functioning poorly in a
number of areas. A higher proportion of
our children had poor-quality relationships
with their caretakers in infancy, as indi-
cated by numbers of infants classified as
anxiously attached at 12 and 18 months
compared to low-risk samples (Egeland
& Farber, 1984). A high proportion had
preschool behavior problems (Erickson,
Sroufe, & Egeland, 1985) and were func-
tioning poorly socially, emotionally, behav-
iorally, and academically in elementary
school (Egeland, & Abery, 1991; Pianta,
Erickson, Wagner, Kreutzer, & Egeland,
1990).
Clearly, poverty was a major risk
condition.
From our findings, the negative effects
of poverty seem to be cumulative and in-
crease as the child gets older. At
12
months,

520
B.
Egeland et al.
62%
of our infants were securely attached,
and at each subsequent assessment the num-
ber of well-functioning children decreased
(Egeland & Kreutzer, 1991). In preschool,
only approximately 30% of the children
were competent in terms of peer acceptance
and cooperation, social skills, emotional
self-regulation, and the ability to play and
function autonomously.
Most children in our sample were
healthy, robust babies who displayed de-
clines in mental, motor, and socioemotional
areas of development. As a group, they
were not born with physical handicaps or
medical problems; however, as a result of
adverse effects of poverty, they showed
maladaptive patterns of development in a
number of areas. Their decline in function-
ing observed at each developmental period
seems to have been related to adverse living
conditions, not inherent factors and traits
within the child. A few children functioned
well despite living in poverty and experienc-
ing a number of adversities. Two of these
children were advanced a grade, and two
students were placed in classes for the gifted
in sixth grade. A number of children were
judged to be well functioning in the social
and emotional areas as well. For example,
15%
of the children were ranked by their
teachers above the 90th percentile on peer
competence/popularity compared to the
other children in the sixth-grade class-
rooms.
Extreme Risk: The Effects of Child
Maltreatment and Resilience
For those children who experienced extreme
risk, especially those who were maltreated
in our sample, we found little evidence of
competence over time. Protective factors
had a relative influence, only diminishing
negative consequences. From an organiza-
tional perspective, it may be expected that
children reared in extremely chaotic home
environments or abusive caregiving rela-
tionships may encounter difficulties in suc-
cessfully resolving early developmental
tasks and, as a result, may be vulnerable to
later deviations or delays in development
(Cicchetti & Schneider-Rosen, 1986).
We identified 44 children during the in-
fancy through preschool period and found
that none of
these
children were functioning
in a competent fashion across this period.
Over 50% were securely attached at 12
months; however, by first grade, none were
judged to be in the competent cluster group
(Farber
&
Egeland, 1987). While some were
performing satisfactory work academically
or were socially skilled, none were compe-
tent in terms of overall adaptation including
both behavior and emotional health.
A few maltreated children showed im-
proved functioning in isolated areas over
the infancy through preschool period. Im-
proved functioning was associated with
placement in foster care (though in some in-
stances, this had a negative effect), the
availability of a caring adult, intensive fam-
ily intervention, changes in family status
(e.g., father not allowed visitation), and a
structured school environment with a car-
ing teacher. Despite periodic improvement
in specific domains, however, adaptation of
these children across time was often poor,
and at the time of the last completed assess-
ment in the sixth grade all of these children
demonstrated clear dysfunction resulting
from maltreatment.
Our longitudinal study of maltreated
children highlights the need for regular as-
sessments of development across time,
broadband measures of adaptation, and a
transactional process view of resilience.
One-time assessments indicating specific
behavioral competencies in isolated do-
mains (e.g., school achievement, aggres-
sion) are insufficient to label children resil-
ient (Farber & Egeland, 1987). Despite
periodic competencies, we found the overall
functioning of maltreated children to be
poor using broad definitions and measures
of adaptation over time. For example, one
of our most severely abused children was
also one of the better functioning children
during the early elementary years; however,
due to changes in her life (removal from a
caring foster family and placement with her
biological mother), her functioning in the

Resilience as process 521
sixth grade deteriorated significantly (she
ultimately was placed in a residential treat-
ment facility).
Continuity in Adaptation and
Factors Related to Discontinuity
A degree of continuity in individual adapta-
tion between infancy and childhood as well
as between childhood and adolescence has
been widely supported in the literature, al-
though change has also been noted (e.g.,
Block, 1987; Mischel, Shoda, & Peake,
1988;
Zahn-Waxler, Mayfield, Radke-
Yarrow, & McKnew, 1988). The degree of
continuity has been found to vary among
longitudinal studies, depending on a variety
of factors including the developmental peri-
ods investigated, age span, and the particu-
lar measures used to assess behavior and
adaptation across time. By using compre-
hensive assessments emphasizing behav-
ioral organization, we have demonstrated
continuity between individual differences in
attachment in infancy and socioemotional
adaptation in preschool (Erickson et al.,
1985) and middle childhood (Sroufe, 1983;
Sroufe, Carlson,
&
Shulman, 1993).
Egeland and Kalkoske (1993) found con-
tinuity between infancy and the early school
years to be a function of
level
of risk and sex
of the child. Using teacher rankings of peer
competence and emotional health and rat-
ings of a variety of emotional and behav-
ioral variables, continuity was found to be
high for low-risk boys compared to that for
boys with multiple risk factors. For girls,
continuity between infancy and the early
school years was low regardless of risk level.
Despite findings supporting considerable
continuity in individual development from
infancy through childhood, an important
focus of our study was the identification of
factors and mechanisms that account for
discontinuity or change in individual adap-
tation. Forty percent of our sample changed
attachment classification between
12
and 18
months (Egeland & Farber, 1984). For
boys,
change in classification from anxious
to secure was related to change in mother-
child interaction, the mother's relationship
status,
and life
stress.
For girls, relationship
and contextual factors appeared to be less
important. Instead, discontinuity was re-
lated to personal characteristics of the
mother.
Also,
we found notable exceptions to a
high degree of continuity in a study of adap-
tation from infancy through the preschool
period (Erickson et al.,
1985;
Sroufe, 1983).
In particular, the quality of the relationship
at
42
months was found to be a strong medi-
ator of adaptation between infancy and the
preschool period. Children who were se-
curely attached in infancy yet had later be-
havior problems had mothers who provided
less emotional support, direction, and
structure in a teaching task activity at 42
months. The quality of the mother-child re-
lationship observed at 42 months also was
found to be related to the change from anx-
ious attachment to competent functioning
in preschool. In addition, family stressful
life events and the quality of the home envi-
ronment accounted for exceptions to pre-
dicted outcomes (Erickson et al., 1985).
Follow-up studies in the early school
years yielded similar results regarding
change (Egeland, Kalkoske, Gottesman, &
Erickson, 1990). In this study, changes in
relationship and contextual factors ac-
counted for discontinuity. Change was re-
lated to mother-child interaction, organiza-
tion of the home environment, the family's
relationship status, family
stress,
and changes
in maternal functioning (e.g., changes in de-
pressive symptomatology). Children whose
functioning worsened across time had expe-
rienced significant decreases in environ-
mental support. Those groups character-
ized by some degree of recovery from earlier
maladaptation ("resilient" groups) had ex-
perienced improved care and/or increasing
support.
Factors Mediating the Effect of
Family Stress and Child Adaptation
For children living in poverty, stressful life
events are numerous and compounded by
adverse social and economic factors. Par-
ents and caregivers serve as mediators of the
effects of poverty, potentially harmful com-
munity values, social isolation, psychosocial

522
B.
Egeland
et al.
pathology, and difficult relationships with
family and societal networks (Musick, Bern-
stein, Percansky,
&
Stott,
1987;
Pianta
&
Ege-
land, 1990). In several studies conducted as
part of
the
Mother-Child Project,
we have
ex-
amined
the
relationships between life stressors
and child outcomes (Egeland & Kreutzer,
1991;
Pianta, Egeland,
&
Sroufe, 1990).
Pianta et al. (1990) found that stressful
family events (experienced during the pre-
school period) were predictive of outcomes
in the socioemotional-behavioral domains
in first grade, even after accounting for ma-
ternal and child IQ. Children exposed to
high levels of maternal stress demonstrated
a range of developmental outcomes. To
identify protective factors, these authors
compared competent to less competent chil-
dren of highly stressed mothers. (Compe-
tent and less competent groups were formed
based on the results of a hierarchical cluster
analysis.) Competent boys of highly
stressed mothers were differentiated from
their incompetent peers by their higher in-
telligence and language ability; the extent to
which their mothers provided a structured,
responsive home environment despite high
stress;
and ratings reflecting positive
mother-child interaction at 42 months old.
The data suggest that competence in boys of
highly stressed mothers was in part due to
the mothers' ability to buffer their sons
from the effects of stress and to continue to
provide their sons with good quality care.
Like the boys, the more competent girls
were more intelligent, had better language
skills,
and lived in more organized home
environments. For girls, however, compe-
tence was most highly related to positive
maternal personality characteristics. It was
suggested that competence in girls may have
depended on their mothers' personal adjust-
ment, which may have had the double bene-
fit of buffering the daughters from the neg-
ative effects of stress and providing a role
model for positive coping.
Egeland and Kreutzer
(1991)
investigated
the protective influence of a history of com-
petent development. High- and low-stress
groups were identified by a median split on
the mothers' stress scores. Children with a
history of competence were identified by ex-
amining their membership in positive adap-
tation or poor adaptation groups formed on
the basis of assessments of mother-infant
attachment at 12 or 18 months, mother-
child interaction in problem-solving situa-
tions at 24 and 42 months, and an individ-
ual problem-solving situation at 42 months.
For boys from high-stress families, a his-
tory of early competence was a strong pro-
tective factor for positive adaptation in the
early school years. A closer examination of
an early history of competence indicated
that the most important protective factors
were secure attachment at
12
and
18
months
and a good quality mother-child relation-
ship observed in the teaching task at 42
months. For girls, a history of competence
during the infancy through preschool pe-
riod was not as strong a protective factor.
These findings, for boys especially, support
an organizational view of resilience as a ca-
pacity that develops over time in the context
of a supportive environment.
Together, the results from these studies
suggest that the relationship between family
stress and child developmental adaptation is
complex and may depend on the child's sex,
age,
developmental capacities, and past his-
tory of developmental adaptation. The rela-
tionship may be mediated by organization
of the home environment, maternal person-
ality traits, and aspects of the mother-child
relationship. Furthermore, parenting and
relationship factors, especially for boys,
seem to provide a protective function for
children whose families are experiencing
large amounts of
stress.
An organizational-
developmental perspective provides a
framework for integrating these complex
relationships. Moreover, such an approach
applied to the study of variations in family
structure and development may further elu-
cidate processes influencing individual ad-
aptation (Cicchetti et al., in press).
Early Experience and Resilience
We documented the role of prior history of
adaptation in later competence in yet an-
other way in the Mother-Child Project

Resilience as process 523
(Sroufe, Egeland, & Kreutzer, 1990). We
compared two groups of children in elemen-
tary school on teacher judgments of peer
competence and emotional health. Children
in both groups had been functioning (equiv-
alently) poorly on three assessments across
the
3
Vi-4'/4-year age period. One group,
however, had shown consistently positive
adaptation during infancy and the toddler
periods whereas the other had functioned
poorly throughout. Children with the early
history of positive transactions within the
caregiving system fared significantly better
in the early school years than did children
with a consistent history of impaired func-
tioning. Positive functioning in the early
school years for the resilient group seemed
to
be
tied to their positive adaptation during
infancy and the toddler period. Early adap-
tation from infancy and the toddler period
continued to influence later adaptation
(early elementary school), even after the ef-
fects of intervening adaptation (preschool)
were taken into account.
These data seemed to support the organi-
zational developmental thesis that current
adaptation is a product of both current
circumstances and developmental history
(Bowlby, 1980). While past and current ex-
perience may contribute independently to
current competence, in other research from
our project (Elicker, Englund, & Sroufe,
1992;
Sroufe & Fleeson, 1988), we have
shown that children differing in early at-
tachment relationships also vary in later re-
lationships with teachers and peers (eliciting
support or rebuff), thus experiencing differ-
ent contemporary environments. The rela-
tionships of
these
factors are complex; how-
ever, the findings seem to support the view
that prior experience may play an enduring
role through expectations and attitudes the
child brings to current experience.
Emotional Responsivity as a
Protective Mechanism
In general, our findings emphasize the im-
portant protective function of parenting, or
the quality of parent-child relationships,
particularly early relationships. These re-
sults are consistent with those from other
investigations, particularly results from
studies of the effects of distal risk factors
such as natural disasters, war, poverty, and
family stressful life events (Garmezy, 1985;
Masten et al., 1990). The findings suggest
that effects of severe risk conditions are me-
diated in part by the care the child receives
from his or her parents or alternative care-
givers (Elder, Nguyen,
&
Caspi, 1985; Mas-
ten et al., 1990; Pianta & Egeland, 1990).
As reported earlier in this paper, we
found that a secure attachment relationship
in infancy serves a crucial protective func-
tion. Crucial for the development of such a
relationship is the experience of sensitive
and emotionally responsive caregiving (Ege-
land & Farber, 1984). We found that chil-
dren who have been chronically deprived of
emotionally responsive caregiving show se-
vere impairment in functioning in all areas
of development (Egeland & Sroufe, 1981).
In the Mother-Child Project,
19
children
with histories of psychologically unavail-
able caregiving were identified in infancy
and the toddler period. As part of a mal-
treatment study, mothers were selected on
the basis of information obtained from ob-
servations of mother and infant in the
home, interviews with mothers, clinic and
laboratory observations, and hospital re-
cords (Egeland
&
Sroufe, 1981). Mothers in
the psychologically unavailable group were
characterized by a lack of responsiveness to
their children and, in many instances, a pas-
sive rejection of them. These mothers were
observed to be generally unresponsive to
their infants' bids for interaction (i.e., in-
fants'
smiles and gaze behaviors as well as
signals of distress). The mothers were de-
scribed as detached, withdrawn, appearing
depressed, and displaying flat affect. At 12
months, 46% of this group of infants was
classified as anxiously attached; by 18
months 700% was classified as anxiously
attached. As toddlers, the children were
rated as noncompliant, angry, and easily
frustrated. They showed a major decline in
scores (38 points) on the Bayley Scales of
Infant Development from 9 to 24 months.
In preschool, these children were described

524
B.
Egeland et
al.
as negativistic, lacking in self-control, and
not well liked by their peers, and they dem-
onstrated a greater number of behavior
problems compared to a control group
from the same high-risk sample (Egeland &
Erickson, 1987; Egeland & Sroufe, 1981).
On various measures of infant tempera-
ment (e.g., nurses, ratings in the newborn
nursery) and individual differences (e.g.,
Brazelton Behavioral and Neurological As-
sessment Scale [Brazelton, 1973] adminis-
tered at Days 5 and 10), the infants of emo-
tionally unresponsive parents did not differ
from the controls. They were constitution-
ally robust babies who showed a steep de-
cline in functioning across the first 6 years
of life. By the time they entered school,
many showed major signs of more severe
psychopathology. For these children, a his-
tory of interacting with an emotionally un-
responsive parent resulted in severe malad-
aptation by school age.
A caring and emotionally responsive re-
lationship was also found to be an impor-
tant factor in explaining why some mothers
were able to break an intergenerational cy-
cle of abuse (Egeland, Jacobvitz,
&
Sroufe,
1988).
In this study, 40% of the mothers
who were abused as children were found to
abuse their children, 30% provided "bor-
derline care," and 30% broke the cycle and
were providing good-quality care. Mothers
who broke the cycle of abuse reported that
foster parents or a relative provided them
with emotional support as a child or adoles-
cent. In addition, a number of these moth-
ers had been involved in long-term therapy
providing them with support and enabling
them to come to acknowledge and integrate
early experiences of abuse into views of self
(Egeland et al., 1988). All found relation-
ships with supportive partners in adult-
hood.
Conclusion
Resilience has been described as the capac-
ity for positive outcomes despite challeng-
ing or threatening circumstances (Masten et
al.,
1990; Rutter, 1990). Most numerous
among longitudinal studies of resilience
among high-risk children are investigations
of the effects of poverty. Elder and his col-
leagues examined the impact of economic
hardship on children who experienced the
Great Depression (Elder, 1974) and found
that boys had more difficulty as children
and adults compared to girls who experi-
enced the Depression. He found that per-
sonal assets such as intelligence (for boys),
physical attractiveness, easy temperament,
and a positive mother-child relationship
were protective factors against the negative
impact of economic hardship. Long and
Vaillant (1984) studied a group of adults
from inner-city Boston who grew up during
the Great Depression and found that the
majority were functioning well. Protective
factors accounting for improved function-
ing in this study, however, included social
and economic conditions (rising economy,
employment opportunities, and postwar oc-
cupational training) strikingly different
from conditions today (Garmezy, 1993).
Werner and Smith (1992) followed 614
children born on Kauai in 1955 through age
31.
Even though over half of the children
were born in poverty, 97% graduated from
high school, only 15% of the boys and 12%
of the girls worked in semiskilled jobs, and
only 1 in 6 marriages ended in divorce.
About half of
the
adults who had been trou-
bled by mental health problems in their
teens were functioning well. The most im-
portant protective factor for the majority of
troubled youth was a supportive spouse.
The resilient adults were described as "loos-
ening old ties that evoked memories of dep-
rivation loss and pain." They were able to
describe the past without rancor and "what
set them apart were life histories that re-
vealed a pattern of gradual mastery, resto-
ration and recovery."
In our longitudinal study of a poverty
sample growing up in the 1980s, we found
the adverse effects of poverty to be cumula-
tive and increasing as the child gets older
and, for children experiencing extreme risk
(maltreatment), little evidence of compe-
tence over time. At the same time and con-
sistent with prior studies, we identified a
range of factors related to resilience includ-

Resilience as process
525
ing an experience of emotionally responsive
caregiving, early competence, an organized
home environment, child capacities of intel-
ligence and language, and a low overall level
of risk.
Some findings regarding risk and resil-
ience in our research, however, differ from
those of previous studies. Although we do
not have outcome data comparable to the
Berkeley and Boston studies of children
who experienced the Great Depression, it
appears that our poverty sample is func-
tioning more poorly than the individuals
who grew up during the Depression or those
studied by Werner and Smith in Hawaii.
One reason for the difference may be the
number of associated risk factors. For ex-
ample, there were fewer single parents and
divorces in the 1930s-1950s than during the
period of our study (1970-1980s). In addi-
tion, our sample included a number of
young mothers, drug-and alcohol-abusing
parents, and much family violence. It ap-
pears that the poor families of the 1980s ex-
perienced different, if not more, overall risk
than poor families of the 1930s-1950s.
Regardless of risk condition, studies of
resilience (Brown, Harris, & Bifulco, 1986;
Cicchetti & Schneider-Rosen, 1986; Ege-
land & Sroufe, 1981; Masten et al., 1990)
consistently highlight the importance of
supportive caregiving in the protective pro-
cess.
In longitudinal studies of children
of depressed parents, Radke-Yarrow and
Sherman (1990) identified children they
called survivors, children who demon-
strated competence despite adverse rela-
tionship and environmental factors. A key
characteristic of survivors was their history
of having received whatever emotional nur-
turance was available in the family. Musick,
Stott, Spencer, Goldman, and Cohler
(1987) also found emotionally responsive
caregiving to be a major protective factor of
young children of mentally ill parents.
From studies based on the Mother-Child
Project, emotionally responsive caregiving
was found to mediate the effects of high-
risk environments (e.g., maternal stress),
especially for boys; to promote positive
change for children with experiences of
poverty and abuse; and to account for inter-
generational change in abusive caregiving
patterns. Early sensitive and emotionally re-
sponsive caregiving, in particular, was
found to promote positive outcomes even
when intervening functioning and condi-
tions were poor.
The protective function of responsive
caregiving experience, particularly early ex-
perience, was delineated from an organiza-
tional perspective by Bowlby (1982) and
others. A primary focus was the role of re-
sponsive caregiving in the context of more
proximal risk experiences (e.g., maltreat-
ment, caregivers with mental illness). Re-
sponsive caregiving is thought to assist the
child in regulating emotional response and
developing confidence in the supportive
presence of
others.
Through repeated inter-
action with a sensitive responsive caregiver,
the child comes to view the self as lovable
and worthwhile and to experience mastery
in the environment. The child develops con-
fidence in the self and ability to elicit posi-
tive responses from others through the de-
velopmental transaction of internal and
external experience in an environmental
context rather than as the result of inherent
traits.
In turn, these early caregiving quali-
ties may serve as protective functions in spe-
cific risk situations (Rutter, 1979) and have
positive influence on later adaptation (Cic-
chetti & Schneider-Rosen, 1986; Sroufe,
1979).
While it has been important to identify
specific factors that pose immediate risk to
individuals or that serve underlying vulner-
ability or protective functions, an organiza-
tional approach to the study of resilience
has provided a means of integrating find-
ings regarding individual risk and protective
factors and focusing on processes of adap-
tation. From such an organizational view,
the capacity for resilience
is
seen as develop-
ing over time through an interaction of con-
stitutional and experiential factors in the
context of
a
supportive environment. Adap-
tation is viewed in terms of successful reso-
lution of the individual tasks most salient
for given developmental periods. We would
argue that definitions and criteria of resil-

526
B.
Egeland
et al.
ience reflect adaptive functioning across
time rather than one-time assessments and
that both manifest behavior and personal
distress by considered in overall assess-
ments of adaptive functioning.
An organizational developmental per-
spective focusing on processes of adapta-
tion in the context of adversity enhances our
study of both normal development and mal-
adjustment (Cicchetti & Schneider-Rosen,
1986;
Sroufe, 1991). The study of resilience
from this view requires the delineation of
processes involved in both adaptive as well
as disordered outcomes. From the study of
disordered outcomes, processes not readily
distinguished in normal samples may be
separated (Sroufe, 1991), and, from studies
of normal adaptation, linkages among as-
pects of social, cognitive, and affective de-
velopment that lead to maladaptive path-
ways may be uncovered (Cicchetti, 1990).
A developmental approach to research
on resilience has clear implications for pre-
vention and intervention. While it is impor-
tant to identify robust factors that foster re-
silience, an organizational approach to this
study enables researchers to address ques-
tions of why and how some individuals are
able to function competently despite over-
whelming odds or in response to traumatic
events. Understanding the processes by
which some individuals remain confident
and develop supportive relationships in the
midst of adversity is crucial to the develop-
ment of effective prevention and interven-
tion strategies (Rutter, 1990). Such studies
may contribute to strategies for intervening
early in development to promote more posi-
tive outcomes for high-risk populations
(Cicchetti
&
Toth", 1992) and to provide de-
velopmentally appropriate methods of aid-
ing individuals who have experienced ex-
treme adversity.
Our findings highlight the importance of
environmental factors in the ongoing trans-
active process. At an early age, this resides
in the quality of the parent-child relation-
ship,
which is greatly influenced by the par-
ents'
capacity to be sensitive and emotion-
ally responsive. These findings have served
as a guide for the development of Project
STEEP, a preventive intervention program
for high-risk parents of young children. The
STEEP program was designed to improve
the parent-infant relationship and enhance
the development of competence in the early
years (Egeland
&
Erickson, 1990). The pro-
gram began for pregnant women who were
about to have their first child. It was com-
prehensive and intensive and had a number
of goals including helping the mother un-
derstand and respond to her infant's cues
and signals. We are encouraged by our pre-
liminary findings and findings from similar
projects (Olds & Henderson, 1989). We are
optimistic that continued study of the pro-
cess of adaptation in high-risk samples will
lead to a better understanding of normal and
pathological development and will have di-
rect relevance for refining existing interven-
tion and prevention programs as well as pro-
viding ideas for developing new programs.
References
Block, J. (1987, April). Longitudinal antecedents of
ego-control
and
ego-resiliency
in late
adolescence.
Paper presented at the biennial meeting of the Soci-
ety for Research in Child Development, Baltimore,
MD.
Bowlby, J. (1980). Loss. New York: Basic Books.
Bowlby, J. (1969/1982). Attachment. New York: Ba-
sic Books.
Brazelton, B. (1973). A neonatal assessment scale.
Philadelphia: J. B. Lippincott.
Brown, G., Harris, T., & Bifulco, A. (1986). Long-
term effects of early loss of parent. In M. Rutter,
C. E. Izard, & P. B. Read (Eds.), Depression in
young
people:
Developmental and clinical perspec-
tives
(pp.
251-296). New York: Guilford Press.
Caldwell, B., & Bradley, R. (1984). HOME
observa-
tion for measurement of the environment. Little
Rock, AR: Center for Child Development and Ed-
ucation.
Cicchetti, D. (1990). An historical perspective on the
discipline of developmental psychopathology. In J.
Rolf,
A. Masten, D. Cicchetti, K. Nuechterlein, &
S. Weintraub (Eds.), Risk and protective factors in
the development of psychopathology (pp. 2-28).
New York: Cambridge University Press.
Cicchetti, D.,
&
Schneider-Rosen, K. (1986). An organi-
zational approach to childhood depression. In M.
Rutter C. Izard,
&
P. R. Read (Eds.),
Depression
in
young
people:
Developmental
and
clinical
perspec-
tives^. 71-134). New
York:
Guilford Press.

Resilience as process
527
Cicchetti, D.,
&
Toth, S.
L.
(1992). Editorial: The role
of developmental theory
in
prevention
and
inter-
vention. Development
and
Psychopathology,
4,
489-493.
Cicchetti, D., Toth, S.
L.,
&
Lynch,
M. (in
press).
The
developmental sequelae
of
child maltreatment:
Im-
plications
of
war-related trauma.
In L. A.
Leavitt
&
N. A. Fox
(Eds.),
Psychological
effects
of
war
and
violence on
children.
Hillsdale, NJ: Erlbaum.
Egeland, B.,
&
Abery, B. (1991). A longitudinal study
of high-risk children: Educational outcomes.
Inter-
national
Journal
of
Disability,
Development and
Education,
38(3),
271-287.
Egeland,
B., &
Erickson,
M. F.
(1987). Psychologi-
cally unavailable caregiving:
The
effects
on
devel-
opment
of
young children
and the
implications
for
intervention.
In M.
Brassard,
S.
Hart,
& B. Ger-
main (Eds.),
Psychological maltreatment
of
chil-
dren
and youth
(pp.
110-120).
New
York: Perga-
mon Press.
Egeland,
B., &
Erickson,
M. F.
(1990). Rising above
the past: Strategies
for
helping
new
mothers break
the
cycle
of
abuse and neglect.
Zero
to
Three,
11(2),
29-35.
Egeland,
B., &
Farber,
E.
(1984). Infant-mother
at-
tachment: Factors related
to its
development
and
changes over
time.
Child
Development,
55(3), 753-
771.
Egeland,
B.,
Jacobvitz,
D., &
Sroufe,
L. A.
(1988).
Breaking
the
cycle
of
abuse. Child Development,
59(4),
1080-1088.
Egeland, B., AKalkoske,
M.
(1993,
March). Continu-
ity as a function
of
risk
status:
Infant
attachment
to
the early
school years. Paper presented
at the
bi-
ennial meeting
of
the Society
for
Research
in
Child
Development,
New
Orleans.
Egeland,
B.,
Kalkoske,
M.,
Gottesman,
N.,
& Erick-
son,
M. F.
(1990). Preschool behavior problems:
Stability
and
factors accounting
for
change.
Jour-
nal
of
Child Psychology
and
Psychiatry, 31(6),
891-910.
Egeland,
B., &
Kreutzer,
T.
(1991).
A
longitudinal
study
of the
effects
of
maternal stress
and
protec-
tive factors
on the
development
of
high risk chil-
dren.
In A.
Greene,
E. M.
Cummings,
& K. Kar-
raker (Eds.),
Life-span
developmental psychology:
Perspectives
on
stress
and
coping
(pp.
61-84).
Hillsdale,
NJ:
Erlbaum.
Egeland,
B., &
Sroufe,
A.
(1981). Developmental
se-
quelae
of
maltreatment
in
infancy.
In R.
Rizley &
D.
Cicchetti (Eds.), New directions for
child
devel-
opment:
Developmental perspectives in child
mal-
treatment.
San
Francisco: Jossey-Bass.
Elder,
G. H.
(1974). Children of the Great Depression.
Chicago: University
of
Chicago Press.
Elder,
G. H., Jr.,
Nguyen,
T. V.,
&
Caspi,
A.
(1985).
Linking family hardship
to
children's lives. Child
Development,
56, 361-375.
Elicker,
J.,
Englund, M.,
&
Sroufe,
L. A.
(1992).
Pre-
dicting peer competence
and
peer relationships
in
childhood from early parent-child relationships.
In
R. Parke &
G.
Ladd (Eds.), Family-peer relation-
ships: Modes
of
linkage
(pp.
77-106). Hillsdale,
NJ: Erlbaum.
Erickson,
M.,
Sroufe,
L. A., &
Egeland,
B.
(1985).
The relationship between quality
of
attachment
and behavior problems
in
preschool
in a
high-risk
sample.
In I.
Bretherton
& E.
Waters (Eds.),
Monographs
of
the Society
for
Research
in
Child
Development,
50(1-2), 147-166.
Farber,
E., &
Egeland,
B.
(1987). Abused children:
Can they
be
invulnerable.
In J.
Anthony
& B.
Cohler (Eds.),
The
invulnerable child
(pp. 253-
288).
New York: Guilford Press.
Garmezy,
N.
(1985). Stress-resistant children:
The
search
for
protective factors.
In J. E.
Stevenson
(Ed.),
Recent
research
in
developmental
psycho-
pathology
(pp.
213-233). Oxford: Pergamon
Press.
Garmezy,
N.
(1993). Vulnerability
and
resilience.
In
D.
C.
Funder,
R. D.
Parke,
C.
Tomlinson-Keesey,
&
K.
Widaman (Eds.),
Studying lives through
time:
Approaches
to
personality
and
development
(pp.
377-398). Washington, DC: American Psychologi-
cal Association.
Long,
J. V. F.,
& Vaillant,
G. E.
(1984). Natural
his-
tory
of
male psychological health, XI: Escape from
the underclass. American Journal
of
Psychiatry,
141,
341-346.
Masten, A.
S.,
Best, K. M.,
&
Garmezy, N. (1990).
Re-
silience
and
development: Contributions from
the
study
of
children
who
overcome adversity.
Devel-
opment
and
Psychopathology,
2,
425-444.
Mischel,
W.,
Shoda,
Y., &
Peake,
P. K.
(1988).
The
nature
of
adolescent competencies predicted
by
preschool delay
of
gratification. Journal
of
Per-
sonality
and Social
Psychology,
54(4),
687-696.
Musick,
J. S.,
Bernstein,
V.,
Percansky,
C,
& Stott,
H. (1987).
A
chain
of
enablement: Using commu-
nity based programs
to
strengthen relationships be-
tween teen parents and their infants.
Zero
to
Three,
8(2),
1-6.
Musick,
J. S.,
Stott,
F. M.,
Spencer,
K. K.,
Goldman,
J.,
&
Cohler, B.
J.
(1987). Maternal factors related
to vulnerability
and
resilience
in
young children
at
risk.
In E. J.
Anthony & B.
J.
Cohler (Eds.),
The
invulnerable
child (pp, 229-252). New York: Guil-
ford Press.
Olds,
D. L.,
&
Henderson,
C. R., Jr.
(1989). The pre-
vention of maltreatment.
In
D. Cicchetti
&
V.
Carl-
son (Eds.), Child
maltreatment:
Theory
and re-
search
on the
causes
and
consequences
of
child
abuse
and
neglect
(pp.
722-763). Cambridge: Cam-
bridge University Press.
Pianta,
R.,
& Egeland,
B.
(1990). Life stress
and par-
enting outcomes
in a
disadvantaged sample:
Re-
sults
of the
Mother-Child Interaction Project.
Journal
of
Clinical Child
Psychology,
19(4),
329-
336.
Pianta,
R.,
Egeland,
B.,
&
Sroufe, A.
(1990).
Maternal
stress
and
children's development: Prediction
of
school outcomes
and
identification
of
protective
factors.
In J. E. Rolf, A.
Masten,
D.
Cicchetti,
K.
Nuechterlein,
&
S.
Weintraub (Eds.), Risk and
pro-
tective factors
in the development
of
psychopathol-
ogy
(VP-
215-235). Cambridge: Cambridge Univer-
sity Press.
Pianta,
R.,
Erickson,
M. F.,
Wagner,
N.,
Kreutzer,
T.,
&
Egeland, B. (1990). Early predictors
of
refer-
ral
for
special services: Child-based measures
ver-
sus mother-child interaction. School Psychology
Review, XIX(2), 240-250.
Radke-Yarrow,
M., &
Sherman,
T.
(1990). Hard
growing: Children
who
survive.
In J. Rolf, A. S.

528
B.
Egelandetal.
Masten, D. Cicchetti, K. H. Nuechterlein, & S.
Weintraub (Eds.), Risk and protective factors in
the development of psychopathology (pp. 97-119).
Cambridge: Cambridge University Press.
Rutter, M. (1979). Protective factors in children's re-
sponses to stress and disadvantage. In M. W. Kent
& J. E. Rolf (Eds.), Primary prevention of psycho-
pathology: Vol. 3. Social competence in children
(pp.
49-74). Hanover, NH: University Press in
New England.
Rutter, M. (1990). Psychosocial resilience and protec-
tive mechanisms. In A.
Rolf,
A. S. Masten, D. Cic-
chetti, K. H. Nuechterlein, & S. Weintraub (Eds.),
Risk and protective factors in the development of
psychopathology (pp. 181-214). New York: Cam-
bridge University Press.
Sroufe, L. A. (1979). The coherence of individual de-
velopment: Early care, attachment, and subse-
quent developmental issues. American Psycholo-
gist,
34, 834-841.
Sroufe, L. A. (1983). Infant-caregiver attachment and
patterns of adaptation