Child Development, May/June 2000, Volume 71, Number 3, Pages 703–706
The Stability of Attachment Security from Infancy to Adolescence
and Early Adulthood: General Discussion
Everett Waters, Nancy S. Weinfield, and Claire E. Hamilton
For over three decades, critics of the developmental and psychometric paradigms have argued that individual
differences are neither stable, coherent, nor clinically significant. The present studies extend a long line of research
demonstrating the coherence of individual development in attachment security. They make it clear that
attachment security can be stable from infancy through early adulthood and that change in attachemnt security
is meaningfully related to changes in the family environment. The task now is to better understand the
roles of cross-age consistency in caregiver behavior and the structure of mental representations of early experience
in stability and change.
INTRODUCTION
According to Bowlby (1973, 1980), experience with
primary caregivers leads to expectations and beliefs
(“working models”) about the self, the world, and relationships.
He described these representations as
persistent and yet open to revision in light of experience.
Persistent attachment representations allow positive
secure base experiences to guide behavior when
someone “stronger and wiser” is not at hand. They
also afford a degree of buffering against future unsupportive
and disappointing relationship experiences.
An unfortunate corollary is that unsupportive care also
results in expectations and beliefs that guide (mis)behavior
and complicate relationships (Bowlby, 1985).
Flexibility in working models is also important.
Openness to experience is a hallmark of Bowlby’s (1969)
control systems motivation model, his view of attachment
development, and his desire for attachment
theory to have a significant impact on clinical practice
(Bowlby 1973, 1988). Bowlby’s emphasis on the importance
of real (as opposed to intrapsychic) events in
personality development and psychopathology was a
major departure from classic psychoanalytic theory.
Although Bowlby (1969) implied that the onset and
consolidation of attachment patterns was accomplished
in early childhood, much of his theory and
clinical work envisioned working models evolving
and responding to experience through adolescence
(Waters & Cummings, 2000).
The results of the three studies presented here
show a mix of continuity and discontinuity in attachment
from infancy to adolescence and early adulthood:1
Two of the studies (Hamilton, 2000; Waters,
Merrick, Treboux, Crowell, & Albersheim, 2000) found
significant continuity over time, and one study
(Weinfield, Sroufe, & Egeland, 2000) found no significant
continuity over time. At first glance, the findings
across these studies might be considered inconsistent.
Closer examination, however, reveals that a commonality
ties these studies together into a coherent picture:
Across the three studies there is consistency in the
role of attachment-related life experiences in marking
continuity and change.
These studies make clear that attachment security
can be stable from infancy through early adulthood
and that change in attachment security is related to
meaningful changes in the family environment. Both
Waters et al. (2000) and Hamilton (2000) found that
the majority of their participants maintained the same
attachment status over time. Waters et al., studying a
middle-class sample, found that changes in attachment
classification were associated with the occurrence
of negative life events. Hamilton, studying an
alternative lifestyle sample, also found that these neg©
2000 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/2000/7103-0016
Summary discussion of Waters, Hamilton, & Weinfield, “The
Stability of Attachment Security from Infancy to Adolescence
and Early Adulthood: General Introduction”; Waters, Merrick,
Treboux, Crowell, & Albersheim, “Attachment Security in Infancy
and Early Adulthood: A Twenty-year Longitudinal Study”;
Hamilton, “Continuity and Discontinuity of Attachment from
Infancy”; and Weinfield, Sroufe, & Egeland, “Attachment from
Infancy to Early Adulthood in a High-Risk Sample: Continuity,
Discontinuity, and Their Correlates.”
1 Although these studies are the only such studies currently
completed in the United States, two other studies of continuity
of attachment from infancy to age 16 have been completed in
Germany. For further information, see Becker-Stoll (1997) and
Zimmermann (1994).
704 Child Development
ative life events operated primarily by maintenance
of already established patterns of insecurity or by
movement from secure to insecure patterns. Overall,
however, there was a moderate rate of attachmentrelated
negative life events in both of these samples.
Weinfield et al. (2000) investigated these same issues
in a highly stressed sample, in which attachment-related
negative life events were far more frequent
and more severe than in the Waters et al. (2000)
and Hamilton (2000) samples. The participants in
this study did not maintain the same attachment
classifications. These were not, however, random
changes; change was associated with specific factors,
such as maternal depression, that have every likelihood
of negatively affecting caregiver availability
and responsiveness.
Rather than being inconsistent, the findings of these
three studies present a coherent picture of attachment
as a dynamic process over the course of development.
Vaughn, Egeland, Sroufe, and Waters (1979) have emphasized
that attachment theory requires both stability
and change. The theory requires significant
stability—but not when existing beliefs about significant
others and relationships are under assault. It
also requires change—but primarily where life experiences
challenge existing beliefs and expectations.
Ainsworth (1995) suggested that research showing a
combination of stability and change requires careful
examination of the rules and processes that govern
both pathways; both stability and change are findings
that merit further exploration.
MODELS AND MECHANISMS
OF STABILITY AND CHANGE
As Waters et al. (2000) indicate, stable individual differences
do not explain behavior; stability is a discovery
that requires an explanation. The present studies
do not implicate any particular models or mechanisms
of stability. A variety of mechanisms might be
in play. Two important candidates are early experience
and consistency in caregiver behavior. Rutter
(1979) has argued persuasively that, although Bowlby
overestimated the risks inherent in separation per se,
gross failures of early care may have long-term effects
on social development. This is especially likely if family
and environmental influences act continuously to
maintain early difficulties (Sameroff & Chandler, 1975;
Waters, Kondo-Ikemura, Posada, & Richters, 1991).
A number of other processes may also be relevant.
For example, Epstein (1973, 1991) suggests that conceptualizations
of self and social relationships formed
early in life are more general and less open to revision
than similar beliefs acquired later in life. Temperament
constructs may also play a role in attachment
stability and change, not as an alternative interpretation
of attachment measures but rather as moderators
of infants’ and children’s responsiveness to negative
life events and insensitive or changing patterns of care.
The role of temperament in the development of adult
attachment representations has yet to be explored. In
addition, developmental theorists have emphasized
that individuals have a significant impact on their environments
(e.g., Plomin, 1989; Scarr & McCartney,
1983). The effects of early experience and also heritable
personality and behavioral traits can exert a continuing
bias on parent–child interaction. The result
can be a stabilizing effect on both the environment
and the child’s individual characteristics. Such processes
are important to understand because they have
important implications for research design and interpretation
and also for prevention and intervention.
Change in attachment classification also requires
explanation, and although these studies demonstrate
that attachment-related life events are associated with
change, no specific process or model is implicated.
The effects of negative life events on attachment security
deserve to be examined in their own right. Negative
events could affect attachment security through a
number of routes. From the point of view of attachment
theory, working models are most likely to change in
response to actual changes in caregiver availability
and responsiveness. For example, marital problems
could produce mood effects or cognitive demands
that interfere with the caregiver’s availability and responsiveness.
Over time attachment representations
might change in response to changes in caregiver behavior.
Of course, negative events do not have to act
directly on the caregiver. They might instead have a
direct impact on another family member and then
spread throughout the family system, thereby interfering
secondarily with caregiving.
Negative events might, also change a child’s expectations
of a caregiver’s availability and responsiveness
directly. This might happen, for example, if a
caregiver becomes chronically ill and the child infers
that he or she is now less available. Attachment representations
might then change before (or without) actual
caregiving failures. Marital discord could have a
similar effect (Cummings & Davies, 1996).
THE ROLE OF DEVELOPMENT IN CHANGE
All three studies presented here involve participants
who are in late adolescence or early adulthood. Therefore,
taking developmental issues into account when
considering possible influences on stability and change
is useful, particularly with respect to the issue of au-
Waters, Hamilton, and Weinfield 705
tonomy from one’s family of origin. Specifically, if
one has experienced attachment-related negative life
events, the presence of developmentally salient autonomy
issues might make change from infant security
to adult insecurity more likely.
If it can be inferred from the answers to the Adult
Attachment Interview questions that an individual
had strong negative experiences or few loving experiences
with caregivers during childhood, a secure classification
can be achieved only by the individual acknowledging
that the experiences were negative and
that they affected adult personality. Strong denial that
negative experiences were actually negative, or denial
that those experiences could have had any effect
on the individual’s development, will lead to a dismissing
classification. A secure classification in the
face of negative experiences requires some rethinking
and analysis of childhood experience. Young adults
who have not gained autonomy from their families of
origin may find it too difficult emotionally and cognitively
to acknowledge and explore poor treatment by
a parent on whom they still depend. It seems plausible,
then, that some shifting may occur later as autonomy
is achieved and there is more opportunity
for psychological exploration of the impact of childhood
experiences.
CONCLUSIONS
The present studies provide descriptive information
about the stability and change of attachment organization
from infancy to late adolescence/early adulthood
in a variety of developmental contexts. They
also provide information about the relation between
negative life events and changes in attachment classifications.
This information is a necessary first step toward
process-level research on attachment stability
and change.
Any of the processes suggested here are consistent
with Bowlby’s view that attachment representations
arise primarily from real experiences rather than intrapsychic
events. Early empirical research on attachment
stability (e.g., Waters, 1978) was undertaken in
response to a situationist critique that claimed individual
differences in attachment were neither stable,
coherent, nor of any practical importance (Masters &
Wellman, 1974; Mischel, 1968). Two decades of research
have demonstrated that, as applied to attachment
security, the situationist critique is incorrect.
This has not come about because situations and environments
proved unimportant. The critique failed
primarily because both stability and change have
proven more complicated and more interesting than
the situationists imagined (Waters et al., 1991).
These studies suggest many possibilities for future
avenues of research. First, there is no longitudinal
data on continuity of AAI classifications from adolescence
to adulthood. Research on AAI stability over
this period of time would lend insight into whether
late adolescent and adult representations are equivalent,
or whether, particularly in cases where there
have been negative attachment-related experiences,
there is sometimes a period of transition for attachment
representations. Second, re-interviewing some
of the participants from the studies presented here as
they complete the transition to adulthood and more
of them become parents would also be informative.
This would allow us to explore whether the rates of
stability remain the same through adulthood. It
would also allow for a prospective examination of intergenerational
patterns of attachment.
Rather than simply resolving questions about continuity
of attachment, the present studies should be
taken as starting points for stimulating more research.
These studies demonstrate that attachment security
can be stable over very long periods of time. They
also demonstrate that high intensity, attachmentrelated
negative events are associated with changes in
attachment security over such intervals. The task now
is to explain the underlying processes.
ACKNOWLEDGMENTS
This work was supported in part by grants to Everett
Waters from The Foundation for Child Development
and to Everett Waters and Judith Crowell from the
National Institutes of Mental Health (MH44935); by
grants to Byron Egeland, L. Alan Sroufe, and W. Andrew
Collins from the National Institutes of Mental
Health (MH40864-08), and by a Dissertation Fellowship
Award granted by the University of Minnesota
to Nancy S. Weinfield; by grants to the late Bernice T.
Eiduson from Carnegie Corporation (B3870) and
USPHS (MH24947) and to Thomas S. Weisner from
Carnegie Corporation (B4189) and W. T. Grant Foundation
(#92-1488); and by dissertation grants to Claire E.
Hamilton from the Psychocultural Study Group, Department
of Psychiatry and Biobehavioral Sciences,
University of California, Los Angeles.
ADDRESSES AND AFFILIATIONS
Corresponding author: Nancy S. Weinfield, Department
of Psychology, Gilmer Hall, University of Virginia,
Charlottesville, VA 22903; e-mail: nweinfield@
virginia.edu. Everett Waters is at the State University
of New York at Stony Brook; and Claire E. Hamilton is
at the University of Georgia, Athens.
706 Child Development
REFERENCES
Ainsworth, M. D. S., & Marvin, R. (1995). On the shaping of
attachment theory and research: An interview with
Mary D. S. Ainsworth (Fall 1994). In E. Waters, B.
Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving,
cultural, and cognitive perspectives on securebase
behavior and working models: New growing
points of attachment theory and research. Monographs of
the Society for Research in Child Development, 60(2–3, Serial
No. 244).
Becker-Stoll, F. (1997). Interaction behavior between adolescents
and mothers in the context of longitudinal attachment development.
Unpublished doctoral dissertation, University of
Regensburg, Germany.
Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment.
New York: Basic Books.
Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation. New
York: Basic Books.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss. New York:
Basic Books.
Bowlby, J. (1985). The role of childhood experience in cognitive
disturbance. In M. J. Mahoney & A. Freeman (Eds.),
Cognition and psychotherapy (pp. 181–200). New York &
London: Plenum Publishing Corporation.
Bowlby, J. (1988). A secure base: Clinical applications of attachment
theory (collected papers). London: Tavistock.
Cummings, E. M., & Davies, P. (1996). Emotional security as
a regulatory process in normal development and the development
of psychopathology. Development and Psychopathology,
8, 123–139.
Epstein, S. (1973). The self revisited or a theory of a theory.
American Psychologist, 28, 404–416.
Epstein, S. (1991). Cognitive-experiential self theory: Implications
for developmental psychology. In M. Gunnar &
L. A. Sroufe (Eds.), Self-processes and development.
Minnesota Symposia on Child Psychology (Vol. 23, pp. 79–
123). Hillsdale, NJ: Erlbaum.
Hamilton, C. E. (2000). Continuity and discontinuity of attachment
from infancy through adolescence. Child Development,
71, 690–694.
Masters, J., & Wellman, H. (1974). Human infant attachment:
A procedural critique. Psychological Bulletin, 81,
218–237.
Mischel, W. (1968). Personality and prediction. New York:
Wiley.
Plomin, R. (1989). Environment and genes: Determinants of
behavior. American Psychologist, 44, 105–111.
Rutter, M. (1979). Maternal deprivation, 1972–1977: New
findings, new concepts, new approaches. Child Development,
50, 283–305.
Sameroff, A., & Chandler, M. (1975). Reproductive risk and
the continuum of caretaking casualty. In. F. Horowitz
(Ed.), Review of Child Development Research (Vol. 4, pp.
187–244). Chicago: University of Chicago Press.
Scarr, S., & McCartney, K. (1983). How people make their
environments: A theory of genotype➝environment effects.
Child Development, 54, 424–435.
Vaughn, B., Egeland, B., Sroufe, L.A., & Waters, E. (1979).
Individual differences in infant–mother attachment:
Stability and change in families under stress. Child Development,
50, 971–975.
Waters, E. (1978). The reliability and stability of individual
differences in infant–mother attachment. Child Development,
49, 483–494.
Waters, E., & Cummings, E. M. (2000). A secure base from
which to explore close relationships. Child Development,
71, 164–172.
Waters, E., Kondo-Ikemura, K., Posada, G., & Richters, J.
(1991). Learning to love: Mechanisms and milestones. In
M. Gunnar & L.A. Sroufe (Eds.), Self processes and development.
Minnesota Symposium on Child Psychology (Vol.
23, pp. 217–255). Hillsdale, NJ: Erlbaum.
Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim,
L. (2000). Attachment security in infancy and early
adulthood: A twenty-year longitudinal study. Child Development,
71, 684–689.
Weinfeld N. S., Sroufe L. A., & Egelund, B. (2000). Attachment
from infancy to early adulthood in a high-risk sample:
Continuity, discontinuity, and their correlates. Child
Development, 71, 695–702.
Zimmermann, P. (1994). Attachment in adolescence: Development
and coping with current demands. Unpublished doctoral
dissertation, University of Regensburg, Germany.