Child Development, May/June 2000, Volume 71, Number 3, Pages 707–720
Attachment over Time
Michael Lewis, Candice Feiring, and Saul Rosenthal
Continuity in attachment classification from infancy to late adolescence was examined and related to autobiographical
memories of childhood, divorce, and maladjustment. Eighty-four White middle-class children (48
girls) were seen in a modified Strange Situation at 12 months and given the Adult Attachment Interview at 18
years. In addition, data were collected on 13-year-olds’ childhood recollections as well as adolescent, mother,
and teacher ratings of maladjustment at 13 and 18 years of age. Divorce status of parents also was obtained.
Results indicated no continuity in attachment classification from 1 to 18 years of age and no relation between
infant attachment status and adolescent maladjustment. Divorce was related to 13-year-olds’ childhood recollections
as well as to insecure attachment status at 18 years. Eighteen-year-olds with insecure attachment classification
were more likely to rate themselves as maladjusted. The results support the idea of attachment as an
evolving representation dependent upon the nature of the family environment as indexed by divorce.
INTRODUCTION
Studying development involves the exploration of
continuity and discontinuity (Lewis & Starr, 1979;
Wohlwill, 1973). The concept of attachment has captured
our attention because it seeks to explain, in part,
the origins of social and emotional behavior. Attachment
theory has generated an enormous amount of
theoretical and empirical work. As a construct, attachment
has been considered both as a set of overt
behaviors (e.g., Ainsworth, Blehar, Waters, & Wall,
1978; Belsky, Rovine, & Taylor, 1984; Waters & Dean,
1982) as well as a representation or model of close relationships
(e.g., Bowlby, 1973; Bretherton, 1985).
These two ways of viewing attachment necessitate a
developmental model that describes how attachment
behaviors in infancy lead to attachment representations
later in life.
Formulations of attachment have usually emphasized
a connection between infants’ early overt behaviors
and later representations (Bowlby, 1969; Sroufe,
1983). Main (1990), for example, believes that the later
models are derived from earlier attachment experiences.
The child’s model of the attachment relationship
is viewed as organized around the history of
the caregiver’s responses to the infant’s actions. Thus,
the construct of attachment as measured by mental
representations is believed to be formed around early
parent–infant experiences. In this view, the Strange Situation
is believed to capture overt behaviors that reflect
an infant’s internal working model of the attachment
relationship. According to Main, Kaplan, and Cassidy
(1985, p. 77); “by the time the infant is one year of age,
individual differences in Strange Situation behavior
with a particular parent may be conceived as reflecting
individual differences in the infant’s internal working
model of a particular infant–parent relationship.”
The construct of secure attachment is measured by
observing sets of behaviors that move the individual
toward and maintain proximity with an attachment
figure. These behaviors are used because infants’ cannot
articulate representations due to their limited cognitive
capacity. In the Strange Situation procedure,
the infant’s proximity and responses to the caregiver
during reunion are used to categorize the type of attachment
relationship (Ainsworth et al., 1978). At
older ages, the need to rely on sets of overt behaviors
becomes less central because models of attachment
can be assessed through individual narratives of attachment
history.
Two different strategies for assessing the attachment
construct—observing overt behaviors and tapping
representations—raise important developmental
questions about continuity of the construct. In the
study of attachment, continuity would be indicated by
a relation between early overt behaviors and later representations.
Although attachment theory argues that
such a relation exists, little empirical work has examined
this continuity. Although we agree that a set of
overt behaviors in early life and later mental models
can represent the same construct, data as to whether
individual differences remain consistent over this developmental
transformation are limited. Findings
from unpublished proceedings of national meetings
are contradictory. One study shows consistency (Waters,
Treboux, Crowell, Merrick, & Albersheim, 1995),
whereas another shows inconsistency (Zimmerman,
Fremmer-Bombik, Spangler, & Grossman, 1997). The
question addressed in this study is whether individual
differences in attachment are consistent over time.
© 2000 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/2000/7103-0017
708 Child Development
Some have suggested that individual consistency
exists over this developmental transformation from
infancy through adolescence. Differences in early
mother–infant interactions produce individual differences
at 1 year of age in sets of overt behaviors
(Ainsworth et al., 1978; Belsky, Rovine, et al., 1984;
Lewis & Feiring, 1989). The differences in types
(A,B,C,D) reflecting the quality of the attachment relationship
are seen as predicting subsequent representations
(Bretherton, 1985). Sroufe (1983; p. 74) argued
that these early overt behaviors serve as a
marker for a prototype that influences the nature of
later intimate relationships. “Even when children
change rather markedly, the shadows of the earlier
adaptation remain, and, in times of stress, the prototype
itself may be clear.”
Studies used to support this supposition have
linked overt attachment behaviors in infancy to later
social competence and psychopathology (e.g., Arend,
Gove, & Sroufe, 1979; Elicker, Englund, & Sroufe, 1992;
Lewis, Feiring, McGuffog, & Jaskir, 1984; Londerville
& Main, 1981; Matas, Arend, & Sroufe, 1978; Pastor,
1981; Sroufe, 1983; Sroufe, Egeland, & Kreutzer, 1990;
Sroufe & Fleeson, 1986). The relations between early
attachment behaviors and later competencies are assumed
to reflect the ongoing impact of a stable attachment
construct. Although these studies show associations
between early overt attachment behaviors and
subsequent competence, they do not examine continuity
in attachment behaviors with later representations;
therefore, they do not provide support for the
idea that infant attachment is a prototype for later attachment.
By assuming that attachment remains stable,
these studies and their interpretation mistakingly
promote the view that infant attachment status is a
trait-like attribute of an individual throughout development.
As Waters, Posada, Crowell, and Lay (1993,
p. 217) have pointed out, “attachment theorists often
referred to infant attachment status as if it were a traitlike
characteristic that an individual carried throughout
life. This has stood both as dogma and doctrine
based on empirical research. Indeed, attachment theorists
are often criticized for offering an ‘inoculation’
theory of development. Secure attachment in infancy
inoculates a child from adverse outcomes throughout
development. Conversely, early attachment difficulties
place the child at risk or even cause subsequent
problems.” The adolescent literature also has treated
attachment as a stable individual characteristic related
to competence and psychopathology (e.g., ColeDetke
& Kobak, 1996; Kobak & Sceery, 1988; Kobak,
Sudler, & Gamble, 1991; Rosenstein & Horowitz, 1996;
Sroufe et al., 1990; Sroufe, Schork, Motti, Lawroski, &
LaFreniere, 1984). Attachment quality is assumed to
be stable and therefore to continue having a significant
impact on intimate relationships and other competencies
into adulthood (e.g., Hazan & Shaver, 1987).
Few longitudinal studies linking attachment in infancy
to attachment in late childhood and adolescence
exist. One central explanation for this paucity of data
relates to measurement issues. Using overt behaviors to
index attachment after early childhood becomes inappropriate.
Although some strategies to assess representations
in childhood exist, they have not been widely
used (e.g., Cassidy, 1988; Fury, Carlson, & Sroufe, 1997;
Oppenheim, Emde, & Warren, 1997). Furthermore, in
the few studies that exist, consistency from overt behaviors
in infancy to behaviors or representations in
childhood is not examined (for exception, see Main et
al., 1985; Zimmerman et al., 1997). No measure exists to
assess representations in early adolescence.
The importance of the social environment is obvious
in the formation of infant attachment quality and
may play a critical role in change in attachment representation
when the child is older. Early mother–infant
interactions are related to the subsequent formation of
1-year quality of attachment (e.g., Ainsworth et al.,
1978; Benn, 1986; Cox, Owen, Henderson, & Margand,
1992; Lewis & Feiring, 1989; for a review, see De Wolff
& van Ijzendoorn, 1997). The question is whether the
caregiving environment affects representations as
they are formed and as they are remembered; the
answer bears directly on the nature of attachment
and whether it changes over time. Bowlby (1982) and
Sroufe (1983) have argued that the nature of the
quality of early attachment will affect the impact of
the caregiving environment on subsequent representations.
Two effects of the ongoing caregiving environment
are possible. In the first, the current representation
is solely dependent upon the current
environment, with the previous representation having
little effect on the current one. Thus, a current secure
environment will produce a concurrent secure
representation regardless of whether the earlier environment
was secure or insecure. In the second
type of effect, an interactive one, the current environment
interacts with the past representation to
produce a current representation.
For either type of effect the nature of the environment
must be taken into account. Some data show
that in early childhood when the caregiving environment
changes, the quality of attachment changes as
well (Belsky, Campbell, Cohn, & Moore 1996; Thompson
& Lamb, 1983a, 1983b). For example, Vaughn,
Egeland, Sroufe, and Waters (1979) found that early
attachment classification changes if maternal behaviors
change, in this case as a function of stressful life
events. To relate early overt behaviors of attachment
Lewis, Feiring, and Rosenthal 709
to subsequent mental representations, the environment
and its change over time must be measured (Bronfenbrenner
& Crouter, 1983; Dunn, 1993; Lewis, 1984;
Suomi, 1979; Waters et al., 1993).
One way to look at the caregiving environment is
to examine the effects of divorce on attachment. Divorce
captures many aspects of the caregiving environment
because it has a direct impact on parents,
children, and the emotional and social experiences in
the family (Davies & Cummings, 1994; Fauber, Forehand,
Thomas, & Wierson, 1990; Grych & Fincham,
1993). As an index of disruption in family life, economic
hardship, and decreased availability of caregivers,
divorce provides the type of environment that
could have an impact on early childhood attachments
as well as effecting change in representation. Moreover,
divorce may serve as an event around which autobiographical
thoughts, feelings, and experiences
with marital conflict are organized in regard to the uncertainty
and negative consequences of love relationships
(Cummings & Davies, 1994; Cummings, Davies,
& Simpson, 1994; Cummings & Smith, 1993;
Davies & Cummings, 1994; Grych & Fincham, 1990).
Children from divorced families have been found to
show more social, emotional, and learning problems
than children from nondivorced families (e.g., Allison
& Furstenberg, 1989; Bray, 1988; Guidbaldi, 1988;
Hetherington & Clingempeel, 1992; Hetherington,
Cox, & Cox, 1982; Zill, 1988; Zimiles & Lee, 1991). Divorce
also has been related to greater likelihood of
having an insecure attachment in 16-year-old adolescents
(Zimmerman et al., 1997).
Another issue in the examination of attachment
consistency is whether current representations of attachment
accurately reflect past representations.
That is, current attachment representations, based in
part on autobiographical memories, may reflect what
actually happened or may be a reconstruction of
what happened (Loftus, 1981). Main and others use
the Adult Attachment Interview (AAI), which relies
on an adult’s ability to integrate specific memories
into a more general understanding of the parent–
child relationship (e.g., Adam, Sheldon-Keller, &
West, 1996; Fonagy, Steele, & Steele, 1991; Kobak,
Cole, Ferenz-Gillies, Fleming, & Gamble, 1993; Main
& Goldwyn, 1991; Main et al., 1985; van Ijzendoorn,
1995; Ward & Carlson, 1995). Through an analysis of
the narrative about attachment-related experiences
with parents and other attachment figures, individual
differences in strategies for regulating the attachment
system are inferred. Although the content and
quality of the autobiographical memories are not the
only criteria from which attachment representations
are inferred, the narratives elicited by the AAI rely on
the adult’s ability and willingness to recount personal
history.
The literature on the construction of autobiographical
memories, although separate from that on attachment,
is relevant to understanding the nature of the
narratives from which inferences are made about representations.Ample
evidence for both adults and children
suggests that autobiographical memories, rather than
reflecting accurate past events, are constructions based
in large part on the individual’s current circumstances
(Schneider & Bjorklund, 1998). In Ross’s (1989) review
of research on memory of one’s personal attributes, he
argues that individuals use their current attributes to
construct their personal histories. He suggests that the
first step in constructing histories involves people noting
their present status; they ask themselves what they
are like now and use this information to make a determination
about the past. Ross (p. 342) argued that
people do this because “the present is generally more salient
and available than a person’s earlier standing.”
Research on the construction of adolescents’ personal
histories indicates that autobiographical memories
are strongly inferred by the adolescent’s current circumstances
and current view of themselves, rather
than verified accounts of earlier childhood events
(Henry, Moffitt, Caspi, Langley, & Silva, 1994). Moreover,
individuals search their memories for events that
are consistent with their current view (e.g., Hamilton,
1981; Hastie, 1981; Loftus, Miller, & Burns, 1978; Taylor
& Crocker, 1981). When asked about their pasts, individuals
are likely to use their current state, circumstances,
and beliefs as an interpretive lens for past experiences.
Using techniques to evoke autobiographical
recollections in children for whom the cognitive demands
of the AAI is inappropriate should provide
information relevant to the examination of representation
consistency across time. In keeping with the literature
on adults, children’s and adolescents’ recollections
of their childhoods would more likely reflect their current
circumstances than reflect what actually happened
(Henry et al. 1994; Yarrow, Campbell, & Burton, 1970).
In the present study, we examine both the consistency
of attachment classification from 1 to 18 years of
age and relations between attachment and autobiographical
recollections at 13 years of age to determine
whether such recollections can inform us as to how
representations remain consistent or change over
time. Because the caregiving environment is likely to
be important in attachment consistency, divorce as a
measure of the caregiving environment was related to
attachment consistency. Finally, we looked at maladjustment
in adolescence as a measure of the child’s
competence as a function of both past and concurrent
attachment representations.
710 Child Development
METHOD
Participants
Participants were 84 one-year-olds recruited for a
longitudinal study from White middle- and uppermiddle-class
suburban families. Fifty-seven percent
of the participants were girls. Parents were initially
contacted regarding the longitudinal study soon after
the birth of their children. The sample was selected to
be balanced for gender and to reflect variations in socioeconomic
status and birth order. The data considered
here were collected at 1, 13, and 18 years of age.
At 18 years, there were 113 participants, of whom 84
had all data at 1, 13, and 18 years. These 84, representing
a 20% attrition rate since infancy, do not significantly
differ from the complete samples within each
age point in the distribution of gender, socioeconomic
status, divorce status, attachment categories, or maladjustment.
Although all families were intact when
participants were 1 year old, by the time they were 18
years of age, 14 of the families, or 17%, had experienced
divorce. This is comparable to the national rate
of 21% during the lifetimes of these children (U.S.
Dept. of Health and Human Services, 1995).
Procedure and Measures
Attachment Classification
One year. Infants and their mothers were seen in
the laboratory in a modified Strange Situation Procedure
within 3 weeks of their first birthdays (see Waters,
Wippman, & Sroufe, 1979, Study 2). A 3-min free play
period, a separation period of 2 min, and a 3-min reunion
period were coded according to procedures
delineated by Ainsworth et al. (1978). In their study,
Waters and colleagues found the distribution of secure
and insecure to be 63% versus 37%. Our use of this
procedure has been reported previously. In the Lewis
et al. study (1984), the distribution of secure (67%)
and insecure (33%) infants agrees with that found for
other middle-class American samples. In the present
study with fewer subjects, the distribution remains
about the same: 71% secure, 29% insecure. Moreover,
in Lewis and Feiring (1989), we reported that mother–
infant interaction at 3 months was significantly related
to both the attachment classification and the
Ainsworth infant avoidance rating scale obtained at 1
year. Similar to Belsky, Garduque, and Hrncir (1984),
Lewis and Feiring (1989) found that moderate levels
of maternal responsivity were associated with secure
attachment whereas high and low levels of responsivity
were associated with insecure attachments. In this
study, for purposes of analysis, the avoidant and anxious
groups were combined so that attachment was a
dichotomous variable coded as insecure or secure.
Eighteen years. Participants were interviewed privately
at home during their senior year of high
school. Interviews lasted approximately 1 hr and followed
the protocol developed by George, Kaplan,
and Main (1985). A young woman conducted all the
interviews. The interviewer asked the adolescents to
describe early relationships with parents, generate
adjectives describing each parent, and provide memories
supporting these adjectives. Adolescents were
asked to recall incidents of distress as a child, including
being upset, separated from parents, rejected by parents,
and threatened by parental separation. They
were then asked to consider the effects of their upbringing
on their personalities; why their parents behaved
as they did; changes in, as well as the current
status of, their relationships with their parents; and
the most important thing learned from the attachment
experiences.
All 84 interviews were transcribed. Two independent
raters, blind to all other measures and trained by
Roger Kobak and Holland Cole, rated the transcripts
by using the Attachment Interview Q-Sort (Kobak et
al., 1993). This Q-Sort assesses both working models
and the organization of thought. Most of the items
were derived from descriptions used in the AAI classification
system (Main & Goldwyn, 1991). One hundred
items indexing the division of security/anxiety
and deactivation/hyperactivation were sorted into
nine categories from most characteristic (9) to least
characteristic (1). The sort had a forced bell-shaped
distribution.1
Following Kobak et al. (1993), the composite QSort
description for each participant was correlated
with the prototype sort for security/anxiety. The
security/anxiety dimension measures the extent to
which a person can flexibly coordinate attachment
with exploration by measuring the parent as a secure
base. Participants with a positive correlation with the
secure prototype are classified secure and those with
negative correlations with the secure prototype are
classified insecure. Kobak et al. reported that classifi-
1 Kobak et al. (1993) used a third rater when the composite reliability
of the first two raters fell below .58 (Spearman-Brown
prophecy formula). Using a third rater for 26% of the cases,
Kobak et al. (1993) reported an average composite reliability of
.74. In this study, a third rater was used for only 12% of the cases,
and the average composite reliability on these transcripts was
.80, with a range of .58 to .93. Although we did not have direct
training in the Main and Goldwyn technique, the use of Kobak
as a consultant and reliability checks between our raters and his
help ensured that our ratings measured the attachment styles
the Q-Sort is intended to measure.
Lewis, Feiring, and Rosenthal 711
cations based on the Q-Sort method, compared with
the classifications derived from the Main and Goldwyn
method, result in 90% agreement for secure and
insecure groups. Research using the Q-Sort technique
suggests that this procedure is applicable for older
adolescents (Kobak & Sceery, 1988).
Autobiographical recollections of childhood. Obtaining
recollections involves asking participants to think
about themselves at earlier ages. For example, Henry
et al. (1994) asked the general question, “Think about
X (. . .) when you were (so many) years old.” At 13
years of age, participants were asked “to think about
and describe their early childhoods.” Before the question,
participants were shown a 1-min videotape of
themselves and their mothers in a positive interaction
in a free-play situation when they were 2 years old, to
prompt memories of their childhoods. Prompting is a
technique that has been found to be useful in directing
a child’s attention toward the past and producing
and sequencing memories (Hudson, 1990). The 13year-olds
viewed this segment alone while their
mothers were busy in another room. Following this
viewing, the experimenter asked them to describe
their early childhoods. After the first statements, the
13-year-olds were given one more opportunity to add
to their initial comments. Each participant’s responses
were video recorded and transcribed.
To score the recollection, a participant’s response
was placed into one of four categories: positive, negative,
mixed, and don’t know. The positive category
reflected a predominance of positive statements but
also could contain some neutral statements (N ϭ 57;
for example, a positive statement was, “My mom and
dad were great. I had a happy childhood”; a neutral
statement, for example, was, “We lived in a big
house.”). The negative category was one where there
were no positive but only negative or neutral statements;
for example, “Got in trouble a lot” or “Confusing,
frightening sort of” (N ϭ 10). The mixed category
contained negative as well as positive and neutral
statements (N ϭ 11). Finally, a don’t know category
was scored for the small number of adolescents who
said “don’t know” or “don’t remember” (N ϭ 6). Two
raters independently scored each teenager’s transcript,
and initial agreement for category membership
was 83%. Disagreements were discussed until a
100% consensus was reached.
Thirteen- and eighteen-year maladjustment. Three
raters—mother, adolescent, and teacher—were used.
Mothers and adolescents completed their ratings of
maladjustment during a laboratory visit. Teacher ratings
were mailed to the school and completed by participants’
English teachers. Comparable forms of the
maladjustment measure were used with mothers, adolescents,
and teachers. Mothers completed the Child
Behavior Checklist (CBCL; Achenbach, 1991a), a standardized
rating scale designed to obtain parents’ reports
of their child’s competencies and behavioral/
emotional problems. Teachers completed the Teacher
Report Form (TRF; Achenbach, 1991b), and the adolescents
completed the Youth Self-Report (YSR;
Achenbach, 1991c). All three instruments are selfadministered
checklists and 89 problem items are common
to all three instruments. Following Achenbach,
maladjustment was defined as a score above the borderline
clinical cutoff points of each instrument (T Ն 60;
Ն85th percentile; Achenbach, 1991a, 1991b, 1991c). In a
nonclinical sample,Achenbach (1991a, 1991b, 1991c) reports
that one should expect to find that approximately
15% of the sample should fall into a maladjusted category,
a similar finding seen in this sample.
RESULTS
Because of our interest in the trajectories of individual
children, we followed an analytic strategy that allowed
us to examine individual children over time.
We used ␹2 analyses because they allowed us to determine
whether there is continuity between attachment,
autobiographical recollections, divorce, and
maladjustment.2
Attachment Consistency
One-year attachment is not related to 18-year attachment
␹2(1, N ϭ 84) ϭ 0.24, ns. Of the insecure
1-year-olds, only 38% (9/24) are insecurely attached
at 18 years of age, whereas for the secure 1-year-olds,
43% (26/60) are insecurely attached. Attachment at
1 year bears no relation to recollections at 13 years, ␹2(3,
N ϭ 84) ϭ 0.76, ns. There is a suggestion that recollections
at 13 years are related to attachment classification
at 18 years, ␹2(3, N ϭ 84) ϭ 6.33, p Ͻ .10. Because
individuals who report only negative recollections
clearly view their childhoods as unhappy periods, we
compared this category with the others. Seventy percent
(7/10) of the adolescents with negative recollections
are insecurely attached at 18 whereas only 37%
(21/57) with positive recall are insecurely attached.
This difference is marginally significant, ␹2(1, N ϭ 67)
ϭ 3.84, p Ͻ .08, as is the difference between adolescents
with negative versus mixed recollections, ␹2(1,
2 Because earlier work on this sample has shown gender differences
in the relation between 1-year attachment and 6-year
maladjustment (Lewis et al., 1984), gender differences were examined
in the relations between attachment at 1 and 18 years
and autobiographical recollections, divorce, and maladjustment.
No gender differences were observed.
712 Child Development
N ϭ 21) ϭ 3.83, p Ͻ .06. Only the negative versus don’t
know groups are not significantly different. Thus, negative
recollections at 13 years appear to be related to insecure
attachment classification at 18 years.
Divorce as an Intervening Variable
Divorce and Attachment
Out of 84 participants, 14 had experienced a family
divorce by the time they were 18 years of age. There
was no relation between attachment at 1 year and
subsequent divorce, ␹2(1, N ϭ 84) ϭ 0.42, ns. No relation
was expected because early attachment status
should not affect later divorce. At 18 years, attachment
classification and divorce are significantly related,
␹2(1, N ϭ 84) ϭ 9.41, p Ͻ .01. Those adolescents
whose parents were divorced are more likely to be
classified as insecure (11/14, or 79%, binomial p Ͻ
.06), whereas those classified as secure are more likely
to be from intact families (46/70, or 66%, binomial
p Ͻ .05).
Because the time of divorce varied, early or later
divorce and 18-year attachment might be related.
Time of divorce was divided into three groups: 2–4
years (N ϭ 5); 5–13 years (N ϭ 4); and 13–18 years
(N ϭ 5). When the time of divorce was compared with
the 18-year-old attachment classification, no relation
was found, ␹2(2, N ϭ 14) ϭ 2.12, ns. How long participants
were in a divorced family was not related to
any of the other measures, so only divorced/not divorced
status is reported.
Divorce and Autobiographical Recollections
Although infant attachment and divorce are unrelated,
there is a significant relation between divorce
and recollection of early childhood, ␹2(3, N ϭ 84) ϭ
11.23, p Ͻ .05. Forty percent (4/10) of the adolescents
in the negative recollections group are from divorced
families, compared with 9% (5/57) for the positive
group and 0% for the mixed (0/11) and don’t know
groups (0/6). Thirteen-year-old adolescents with
negative recollections are significantly more likely to
come from families that are divorced than all other
13-year-olds, ␹2(1, N ϭ 84) ϭ 10.18, p Ͻ .05.
Divorce as a Moderator between Early
and Later Attachment
Bowlby (1973) conceived of developmental pathways
to adjustment or problems as the continuous
branching of railroad tracks. Because of our interest in
individual trajectories of attachment coherence and relations
with autobiographical recollections and divorce,
we followed an analytic strategy that allows us to examine
individual children’s paths over time. We determined
the path each individual followed in terms
of attachment and divorce. Table 1 presents the distribution
of attachment classification at 18 years as a
function of divorce and attachment at 1 year. Presentation
of the data in this form allows for consideration
of individuals and, at the same time, allows the interaction
of early attachment and divorce on later attachment.
At the bottom of Table 1, eight attachment
groups at 18 years are presented as a function of both
earlier attachment and divorce status. For example,
three adolescents who are insecurely attached at 18
years were from divorced families and were insecurely
attached at 1 year (Group 1). Likewise, 33 adolescents
who are securely attached at 18 years are
from intact families and were securely attached at 1
year (Group 8). As already shown, attachment at 18
years is related to divorce status but not to attachment
Table 1 Attachment at 18 Years as a Function of Attachment at 1 Year and Family Divorce Status
Attachment at 1 Year
Insecure Secure
24 60
Divorce Status at 18 Years
Divorced Not Divorced Divorced Not Divorced
5 19 9 51
Attachment at 18 Years
Insecure Secure Insecure Secure Insecure Secure Insecure Secure
3 2 6 13 8 1 18 33
Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 Group 8
Lewis, Feiring, and Rosenthal 713
in infancy. The question now posed is whether there
are any interactions between attachment in infancy
and family divorce as they affect attachment at 18
years. First consider the results for those individuals
who were securely attached at 1 year (N ϭ 60). Securely
attached 1-year-olds of divorced families show
significantly more insecure attachment at 18 years
than securely attached 1-year-olds of intact families,
Groups 5 and 6 versus Groups 7 and 8; ␹2(1, N ϭ 60) ϭ
8.95, p Ͻ .05 with Bonferroni adjustment. Eight out of
9, or 89%, of the securely attached 1-year-olds from
divorced families show insecure attachment at 18
years of age (binomial p Ͻ .05). In contrast, only 18 of
51, or 35%, of securely attached 1-year-olds who come
from intact families show insecure 18-year attachment.
Thus, secure attachment at one year does not
buffer children from developing insecure attachments
at 18 years if they are from divorced families.
Looking at the 24 insecurely attached 1-year-olds,
a similar pattern emerges. Although not significant,
insecure 1-year-olds from families who divorce
(Groups 1 and 2) are more likely to be insecurely attached
at 18 years (3/5, or 60%), whereas insecure
1-year-olds whose families did not divorce (Groups 3
and 4) are less likely to be insecurely than securely attached
at 18 years of age (6/19, or 32%).
Of particular interest is the comparison between
insecure and secure 1-year-olds whose families divorced.
If attachment at 1 year and divorce interact in
their contribution to attachment at 18 years, one
would expect that adolescents who were insecure at
1 year and come from divorced families (Groups 1
and 2) should be more likely to be insecure at 18 than
adolescents who were secure at 1 year and come from
divorced families (Groups 5 and 6). This is not the
case, ␹2(1, N ϭ 14) ϭ 1.59, ns. Sixty percent of adolescents
who were insecure at 1 year and from divorced
families are insecure at 18 years, whereas 90% of adolescents
who were secure at 1 year and from divorced
families are insecure at 18 years.
It also would be expected that the adolescents who
were secure at 1 year and from intact families (Groups
7 and 8) should be more likely to be secure at 18 years
than the adolescents who were insecure at 1 year and
from intact families (Groups 3 and 4). Again, this is
not the case, ␹2(1, N ϭ 70) ϭ 0.08, ns. Sixty percent of
adolescents who were secure and from intact families
are secure at 18 years, whereas 68% of adolescents
who were insecure at 1 year and from intact families
are secure at 18. Thus, there appears to be little interaction
between 1-year attachment and divorce as it
impacts on the 18-year attachment classification.
Although 1-year attachment and autobiographical
recollections of childhood were unrelated, divorce
status at 13 years could interact with 1-year attachment
to affect recollections at 13 years of age. However,
analyses of the interactions showed no effects.
Adjustment and Attachment
Although earlier attachment is unrelated to later
attachment, earlier attachment may be related to later
maladjustment. As expected, there was little agreement
between adolescents, mothers, and teachers in
regard to which 13- or 18-year-olds are maladjusted.3
Because of this lack of agreement between raters, the
relation between attachment and subsequent maladjustment
is considered for each rater and across all
raters.
Adolescents were scored as maladjusted if they
had a rating above clinical cutoff scores. Relations between
early attachment and later ratings of maladjustment
were tested by using ␹2 analyses. The results
show that insecure attachment at 1 year is not related
to adjustment at 13 years as rated either by the adolescents
themselves or their mothers. For those adolescents
who rated themselves as having problems at
13 years, only 14% (2/14) were insecure at 1 year with
the majority of maladjusted adolescents coming from
the secure group (86%). Similarly, for those adolescents
whose mothers rated them as being maladjusted,
only 25% (2/8) were from the insecure group
at 1 year with 75% of the maladjusted adolescents
coming from the secure group. For teachers, there is a
significant association between attachment at 1 year
and ratings of child maladjustment at 13 years, ␹2(1,
N ϭ 81) ϭ 6.14, p Ͻ .01. Contrary to expectation, all of
the 13-year-olds rated as maladjusted by their teachers
were secure 1-year-olds (13/13). Rather than looking
at the ratings of any particular individual, asking
whether any one or more raters rated the child as maladjusted
may be more useful (Fendrich, Weissman,
& Warner, 1991). Attachment at one year was compared
with whether at least one person rated the adolescent
as maladjusted. Analyses indicated a significant
relation, such that securely attached infants are
more likely to be rated as maladjusted at 13 years of
age, ␹2(1, N ϭ 84) ϭ 5.31, p Ͻ .05.
There are no significant relations between insecure
attachment at 1 year and ratings of 18-year-old maladjustment.
Only 18% (5/28) of 18-year-olds who
rated themselves as maladjusted were insecurely attached
at 1 year. For mothers’ ratings it was 57% (4/7),
3 At 13 years, mother and adolescent agreement on maladjustment
was ␬ ϭ .00; teacher and adolescent, ␬ ϭ .00; mother and
teacher, ␬ ϭ .07. At 18 years, mother and adolescent agreement was
␬ ϭ .11; teacher and adolescent, ␬ϭ .14; mother and teacher, ␬ ϭ .27.
714 Child Development
a marginally significant finding, ␹2(1, N ϭ 84) ϭ 3.05,
p Ͻ .10. For the teachers’ ratings, only 33% (2/6) of
maladjusted adolescents had been insecure at one
year. There is no significant relation between attachment
at 1 year and whether at least one person rated
the 18-year-old as maladjusted.
Although attachment at 1 year may not be related
to later maladjustment, concurrent attachment might
be related to concurrent maladjustment. Table 2
shows the results for the relations between concurrent
attachment and maladjustment ratings at 18
years. For adolescents’ ratings, maladjustment is associated
with an insecure classification, p Ͻ .01. Sixtyfour
percent of the adolescents who rated themselves
as maladjusted are from the insecurely attached
group, binomial p Ͻ .05. Teachers’ ratings of maladjustment
are marginally associated with insecure attachment,
p Ͻ .08. Eighty-three percent of the adolescents
who are rated by their teachers as maladjusted
are from the insecurely attached group. For mothers’
ratings of maladjustment, there is no significant effect,
although 71% of the adolescents rated by their
mothers as maladjusted are from the insecurely attached
group. When the criterion of at least one person
rating maladjustment is used, the findings are consistent
with results using individual raters. Adolescents
who are rated by at least one person as maladjusted
are currently more likely to have an insecure attachment
classification, p Ͻ .01.
Adjustment and Autobiographical Recollections
Recollections at 13 years may be related to maladjustment.
Results of ␹2 analyses indicate no concurrent
relations between the recollections of childhood
and any of the raters’ judgments of maladjustment at
13 years. The relation between recollection at 13 years
and maladjustment ratings at 18 yields some results
of interest. Those 13-year-olds who recall their childhoods
negatively are more likely to rate themselves as
maladjusted (60%, 6/10) than those who do not recall
their childhoods as negative (30%, 22/74), a marginal
trend, ␹2(1, N ϭ 84) ϭ 3.63, p Ͻ .07. This pattern shows
a significant effect when at least one person rated the
adolescent as maladjusted, ␹2(1, N ϭ 84) ϭ 4.49, p Ͻ
.05. Seventy percent (7/10) of adolescents who reported
negative recollections are rated as maladjusted
by at least one of the raters, whereas only 35% (26/74)
who are in the positive, mixed, or don’t know groups
are rated as maladjusted.
Divorce and Adjustment
Eighteen-year-old adolescents’ ratings of their
maladjustment are significantly related to divorce,
␹2(1, N ϭ 84) ϭ 7.24, p Ͻ .05. Nine out of 14, or 64%,
adolescents from divorced families rate themselves as
maladjusted, whereas only 19 out of 70, or 27%, adolescents
from intact families rate themselves as maladjusted.
For mothers’ ratings of their adolescents,
29% (4/14) of those who have divorced rate their adolescents
as maladjusted, whereas only 4% (3/70) of
those from intact families rate their adolescents as
maladjusted, ␹2(1, N ϭ 84) ϭ 9.01, p Ͻ .05. Teachers’
ratings show no relation to divorce status. Having at
least one person rating the 18-year-olds as maladjusted
is related to divorce status, ␹2(1, N ϭ 84) ϭ 7.28,
p Ͻ .05. Of the adolescents from divorced families,
71% (10/14) are rated as maladjusted by at least one
rater, whereas only 33% (23/70) of adolescents from
intact families are rated as maladjusted.
DISCUSSION
The results of this study bear on several important issues:
(1) the continuity between attachment behaviors
and attachment representations in adolescents,
(2) the relation between autobiographical recollections
of childhood and attachment representations,
(3) the impact of divorce on attachment representations
and autobiographical recollections, and (4) atTable
2 Relations between Attachment at 18 and Maladjustment at 18 Years
Raters
Adolescents Mothers Teachers Number of Raters
Adjusted Maladjusted Adjusted Maladjusted Adjusted Maladjusted 0 1 or More
Secure 39 10 47 2 42 1 37 12
Insecure 17 18 30 5 26 5 14 21
␹2 ␹2(1, N ϭ 84) ϭ 8.84* ␹2(1, N ϭ 84) ϭ 2.78 ␹2(1, N ϭ 81) ϭ 4.61ϩ ␹2(1, N ϭ 84) ϭ 10.79**
* p Ͻ .05; ** p Ͻ .01; ϩ p Ͻ .10.
Lewis, Feiring, and Rosenthal 715
tachment and maladjustment. The limitations of this
longitudinal study and a conceptualization of changes
in attachment follow.
Our findings reveal a lack of continuity between
overt attachment behaviors at 1 year of age and representations
in adolescence. Others also have found a
lack of continuity from infancy, both for predicting attachment
behaviors in childhood (Belsky, Spritz, &
Crnic, 1996; Thompson & Lamb, 1983a, 1983b) and for
predicting representations in adolescence (Zimmerman
et al., 1997). Even when consistency is found, significant
numbers change their attachment status over time
(Main et al., 1985; Waters et al., 1995). This suggests that
continuity in attachment may be influenced by factors
such as family stress and environmental risk.
Although there are few published studies on continuity,
the belief that attachment classification should be
stable remains strong (Ainsworth & Marvin, 1995). The
belief rests on the idea that attachment representation is
outside of consciousness and resistant to change
(Bowlby, 1980; Bretherton, 1985). A recent view of attachment
emphasizes the importance of ongoing experiences
for the continuity of attachment representations
(Sroufe, 1997). This is consistent with longitudinal research
showing that environments play a central role in
the continuity of individual adaptation (Sameroff,
Bartko, Baldwin, Baldwin, & Seifer, 1988).
It is recognized that the caregiving environment,
particularly maternal behavior and characteristics, is
related to continuity of the attachment classification
(Egeland & Farber, 1984; Egeland & Sroufe, 1981;
Lamb, Thompson, Gardner, & Charnov, 1985). Looking
at the continuity of early overt behaviors in infancy
to representations in adulthood, Waters et al.
(1995) found that individuals who experienced major
negative life events such as severe illness, loss of a
parent, or parental divorce were more likely to show
discontinuity. Thus data showing that environment
has impacts on continuity of attachment do exist. We
also found that divorce is related to discontinuity between
attachment behaviors in infancy and representations
in adolescence; adolescents experiencing parental
divorce were more likely to show insecure
attachment when they were 18 years of age, regardless
of their attachment status at 1 year.
Similar to others examining the continuity of attachment
into adolescence (Zimmerman et al., 1997),
we found that divorce relates to an insecure representation
of attachment. Our findings on divorce are consistent
with studies showing that marital conflict is
related to insecure attachments (Cox & Owen, 1993;
Howes & Markman, 1989; Waters et al., 1993). Divorce
represents a lack of parental availability and an
increase in negative interactions between parent and
child, thus increasing the likelihood of an insecure attachment.
In addition, divorce may provide the child
with a model for showing that close relationships are
not to be counted on (Cummings & Davies, 1994;
Fiese, Wilder, & Bickham, in press; Harold & Conger,
1997; Hetherington et al., 1982). The negative impact
on attachment relationships of a disruption in the
availability of attachment figures has long been argued
(Bowlby, 1973).
The environment appears to influence the changing
representation of attachment. The influence of the
environment can itself affect the new representation,
or the environment in interaction with the old representation
can affect the new representation. When we
examined the interaction of 1-year attachment classification
and divorce on 18-year attachment, there was
no evidence that divorce moderates lawful changes.
For example, secure infants who experienced divorce
were just as likely to be insecure as adolescents as
were insecure infants who experienced divorce. Thus,
early secure attachment status did not buffer the negative
effect of divorce on adolescent attachment.
In this study, representations were obtained through
the observation of attachment behaviors and through
interview procedures. The literature shows that representations
and memories are based in part on current
status (Ross, 1989). This suggests that the recall of past
relationships, when the standard attachment interview
is unavailable, may be useful. Bretherton (1985) has argued
that the evaluation of childhood should be related
to attachment representations. In addition, Main
et al. (1985) have argued that theAAI uses recollections
of memories, although it is not solely dependent on
them. Thus, observing the relation between childhood
recollection at 13 and attachment at 1 and 18 years was
useful. Although no relation was found between recollections
at 13 years and attachment at 1 year, recollections
were related to attachment representations at 18
years. Thirteen-year-old adolescents with negative recollections
of their childhoods were more likely to have
insecure attachment representations at 18 years of age.
Thus, continuity from early adolescent autobiographical
memories of childhood to attachment representations
in late adolescence has some support. Similarly,
other work has shown continuity in children’s representations
of mothers as unavailable and subsequent
insecure attachment representations in adolescence
(Zimmerman et al., 1997).
The change in attachment, as referenced by overt
behavior at 1 year and representation at 18 years, is
influenced by divorce, a marker variable for a range
of processes and events. Divorce can be considered as
an event or as a status variable. For example, emotional
and social disruptions associated with parental
716 Child Development
conflict and fathers leaving may be viewed as an
event like any other trauma. Alternatively, divorce
can be considered both as an event and as a status
variable. The status of being from a divorced family
continues over time; that is, living in a family of only
a mother and interacting with the nonpresent father.
(In our study, with a cohort that was born 1971–1972,
mothers maintained primary custody of the children.)
Our data show no effects of when the divorce took
place. This suggests that being in a divorced family is
an important ongoing status variable. Thus, while divorce
took place earlier than 18 years, its impact on the
nature of the family environment continues. Children’s
recollections at 13 years bear no relation to their
early attachment classification nor to the interaction
between earlier attachment and divorce but are influenced
by divorce status. Thirteen-year-olds who are
from divorced families are significantly more likely to
have negative recollections of childhood. This may be
a function of several factors. There are the ongoing disruptions
in family life related to divorce such as living
in the separate households of each parent and dealing
with step parents and siblings. In addition, negative
emotions aroused by divorce and subsequent changes
in the family can predispose children to evaluate their
childhoods negatively and recall negative experiences
(Grych & Fincham, 1990).
A history of linking early attachment to later competence
has grown out of Bowlby’s (1973) belief that
insecure attachment is a risk factor for psychopathology
(e.g., Bowlby, 1973; Cassidy & Kobak, 1988; Sroufe,
1983, 1997). In this study, we found that attachment at
1 year did not predict maladjustment in adolescence.
Rather, it is divorce that is related to later maladjustment.
Children from divorced families show greater
maladjustment than do children from intact families.
As was true for the relation between early attachment
and later attachment, we failed to find the expected relation
between early attachment behaviors and subsequent
functioning. Furthermore, we found no evidence
for an interaction between attachment at 1 year
and divorce status on maladjustment in adolescence.
Although there is no predictive relation of early attachment
to maladjustment, there is evidence for a concurrent
association. Children with negative recollections
at 13 years are more likely to show maladjustment
at 13 years. Likewise, adolescents who are insecure at
18 years are more likely to show maladjustment at 18
years. These findings are consistent with other work
showing that individuals classified as insecurely attached
are more likely to show symptoms of maladjustment
(Cole-Detke & Kobak, 1996; Fonagy et al.,
1996; Kobak & Sceery, 1988; Kobak et al., 1991; Rosenstein
& Horowitz, 1996).
Limitations of this study must be considered, especially
the measurement of attachment. Given a study
of 18 years in length, changes in methodologies used
to measure constructs are inevitable. For example, the
modified procedure to assess attachment at 1 year differs
from the standard Strange Situation. A number of
factors, however, support the validity of the modified
procedure. First, Waters et al. (1979) used and validated
an identical procedure. Second, as described by
Lewis et al. (1984), the distribution of secure, insecureavoidant,
and insecure-ambiguous middle-class infants
is similar to distributions found in studies using
the unmodified procedure. Moreover, infants’ classifications
derived from the modified procedure predict
later sociability and childhood maladjustment in
ways that are consistent with attachment theory
(Lewis & Feiring, 1989, 1991; Lewis et al., 1984).
Our rating of attachment classification for the 18year-olds
uses the Kobak et al. (1993) Q-sort method.
Kobak et al. reported that the percentages of secure
and insecure 18-year-olds are consistent with those
found by using Main and Goldwyn’s (1991) procedure
for analyzing the Adult Attachment Interview.
Even though we were not trained in the Main and
Goldwyn system, our ratings are consistent with
Kobak’s. Thus, using the Q-sort coding system for the
adult attachment interview is unlikely to be the major
source of difficulty in failing to find continuity from
overt behaviors in infancy to representations in adolescence.
The data on divorce and maladjustment indicate
that the attachment classification at 18 is coherent
with maladjustment as rated by child, teacher, or
mother and also is related to divorce.
The measure of 13-year-old autobiographical recollections
was obtained over 12 years ago, before the
existence of any standardized measure of attachment
for this age group. Furthermore, the cognitive demands
of the adult attachment interview and the nature of the
coding system are inappropriate for young adolescents.
Therefore, we did not have a measure of attachment
representation in early adolescence. Instead, we
used a procedure that is designed to elicit an evaluative
representation of childhood by prompting for
memories (Hudson, Fivush, & Kuebli, 1992). Although
such a procedure is not a direct measure of attachment
representation, it does use memories in common
with the more standard procedures. Support for
the idea that the method is a useful one comes from
some evidence of consistency of these autobiographical
recollections with attachment representations at
18 years of age.
Although measurement issues cannot be dismissed,
the data as reported here are coherent. Attachment at
1 year of age as measured by overt behaviors does not
Lewis, Feiring, and Rosenthal 717
show continuity with attachment at 18 years of age as
measured by representations. However, concurrent attachment,
regardless of the age or type of assessment,
is related to the nature of the caregiving environment.
Previous work has shown that there is a relation between
overt attachment behaviors and mother–child
interactions (DeWolff & van Ijzendoorn, 1997). At 18
years, attachment classification is related to concurrent
familial environment; that is, it is related to whether
the parents are or are not divorced when the child is 18
years old. In addition, attachment is related to concurrent
maladjustment. Moreover, adult attachment is related
to concurrent parenting behaviors (Ainsworth &
Eichberg, 1991; Main et al., 1985; Posada, Waters,
Crowell, & Lay, 1995; van Ijzendoorn, 1995; van Ijzendoorn,
Kranenburg, Zwart-Woudstra, von Busschbach,
& Lambermon, 1991). Together, these findings
suggest that although the attachment representation is
not stable over time, current representations are related
to important indicators of current functioning.
The early studies of attachment did not present
any data on the environment in relation to the outcome
measures. Without such data, we cannot address
whether the consistency of the environment or
characteristics of the child accounts for continuity.
The few studies that include some indicators of the
caregiving environment (Vaughan, Egeland, Sroufe,
& Waters, 1979; Waters et al., 1995) reveal that continuity
in attachment is affected by environmental factors.
Although attachment theory has long emphasized
the importance of the attachment construct as a
continuous developmental force throughout the
lifespan, recently Sroufe (1997) has argued that ongoing
experiences interact with attachment relation history.
As Waters et al. (1993) have argued, emphasizing
the continuity of the caregiving environment should
provide a better foundation for attachment theory
than emphasizing transcendent effects of early experience.
Whereas early insecure attachment may potentiate
short term deviant development (Lewis et al.,
1984; Main et al., 1985; Sroufe et al., 1990), there is no
evidence for an enduring effect. The data presented in
this study support and further this conceptualization
of attachment theory. Stated simply, what is happening
concurrently in the child’s environment (the divorce
status of the family) at least through adolescence,
exerts powerful effects on the child’s attachment behaviors
and representations. The degree to which environmental
factors remain consistent represents the
degree to which the individual’s attachment remains
consistent. Our findings fit with contextual models of
development (James, 1895/1975; Pepper, 1942). A contextual
model emphasizing concurrent adaptation, in
this case to a divorced family life, is a central premise
for understanding children’s emotional and social development.
Such a view is supported by the developmental
literature (Lerner, 1986; Lewis, 1997).
ACKNOWLEDGMENT
This work was supported by a grant from the William T.
Grant Foundation.
ADDRESSES AND AFFILIATIONS
Corresponding author: Michael Lewis, Institute for
the Study of Child Development, UMDNJ-Robert
Wood Johnson Medical School, 97 Paterson Street,
New Brunswick, NJ 08903; e-mail: lewis@umdnj.edu.
Candice Feiring and Saul Rosenthal were at the Institute
for the Study of Child Development, UMDNJRobert
Wood Johnson Medical School, New Brunswick,
NJ at the time of this study. Dr. Feiring is currently at
UMDNJ-New Jersey Medical School, Newark, NJ; and
Dr. Rosenthal is currently at the Massachusetts School of
Professional Training, Boston, MA.
REFERENCES
Achenbach, T. M. (1991a). Manual for The Child Behavior
Checklist and 1991 Child Behavior Profile. Burlington: University
of Vermont, Department of Psychiatry.
Achenbach, T. M. (1991b). Manual for the Teacher’s Report
Form and 1991 TRF Profile. Burlington: University of Vermont,
Department of Psychiatry.
Achenbach, T. M. (1991c). Manual for the Youth Self-Report
and 1991 YSR Profile. Burlington: University of Vermont,
Department of Psychiatry.
Adam, K. S., Sheldon-Keller, A. E., & West, M. (1996). Attachment
organization and history of suicidal behavior
in clinical adolescents. Journal of Consulting and Clinical
Psychology, 64, 264–272.
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S.
(1978). Patterns of attachment: A psychological study of the
Strange Situation. Hillsdale, NJ: Erlbaum.
Ainsworth, M. D. S., & Eichberg, C. G. (1991). Effects on
infant-mother attachment of mother’s unresolved loss of
an attachment figure or other traumatic experience. In
C. M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment
across the life cycle (pp. 160–183). New York and
London: Tavistock/Routledge.
Ainsworth, M. D. S., & Marvin, R. S. (1995). On the shaping
of attachment theory and research: An interview with
Mary D. S. Ainsworth. Monographs of the Society for Research
in Child Development, 60(2–3, Serial No. 244).
Allison, P. D., & Furstenberg, F. F. (1989). How marital dissolution
affects children: Variations by age and sex. Developmental
Psychology, 25, 540–549.
Arend, R., Gove, F. L., & Sroufe, L. A. (1979). Continuity of
individual adaptation from infancy to kindergarten: A
718 Child Development
predictive study of ego-resiliency and curiosity in preschoolers.
Child Development, 50, 950–959.
Belsky, J., Campbell, S. B., Cohn, J. F., & Moore, G. (1996). Instability
of infant-parent attachment security. Developmental
Psychology, 32, 921–924.
Belsky, J., Garduque, L., & Hrncir, E. (1984). Assessing performance,
competence, and executive capacity in infant
play: Relations to home environment and security of attachment.
Developmental Psychology, 20, 406–417.
Belsky, J., Rovine, M., & Taylor, D. G. (1984). The origins of
individual differences in infant-mother attachment: Maternal
and infant contributions (The Pennsylvania Infant
and Family Development Project). Child Development, 55,
718–728.
Belsky, J., Spritz, B., & Crnic, K. (1996). Infant attachment security
and affective-cognitive information processing at
age 3. Psychological Science, 7, 111–114.
Benn, R. K. (1986). Factors promoting secure attachment relationships
between employed mothers and their sons.
Child Development, 57, 1224–1231.
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, sadness,
and depression. New York: Basic Books.
Bray, J. H. (1988). Children’s development during early remarriage.
In E. M. Hetherington & J. D. Arastech (Eds.),
Impact of divorce, single parenting, and stepparenting on children
(pp. 279–298). Hillsdale, NJ: Erlbaum.
Bretherton, I. (1985). Attachment theory: Retrospect and
prospect. In I. Bretherton & E. Waters (Eds.), Growing
points of attachment theory and research. Monographs of
the Society for Research in Child Development, 50(1–2, Serial
No. 209).
Bronfenbrenner, U., & Crouter, A. C. (1983). The evolution
of environmental models in developmental research. In
W. Kessen & P. H. Mussen (Eds.), History, theory, and
methods: Handbook of child psychology (Vol. 1, pp. 357–
414). New York: Wiley.
Cassidy, J. (1988). Child-mother attachment and the self in
six-year-olds. Child Development, 59, 121–134.
Cassidy, J., & Kobak, R. (1988). Avoidance and its relation to
other defensive processes. In J. Belsky & T. Neworski
(Eds.), Clinical implications of attachment (pp. 300–323).
Hillsdale, NJ: Erlbaum.
Cole-Detke, H., & Kobak. R. (1996). Attachment processes
in eating disorder and depression. Journal of Consulting
and Clinical Psychology, 64, 282–290.
Cox, M. J., & Owen, M. T. (1993, March). Marital conflict and
conflict negotiation: Effects on infant-mother and infantfather
relationships. In M. Cox & J. Brooks-Gunn
(Chairs), Conflict in families: Causes and consequences.
Symposium conducted at the meeting of the Society for
Research in Child Development, New Orleans, LA.
Cox, M. J., Owen, M. T., Henderson, V. K., & Margand, N. A.
(1992). Prediction of infant-father and infant-mother attachment.
Developmental Psychology, 28, 474–483.
Cummings, E. M., & Davies, P. T. (1994). The impact of parents
on their children: An emotional security perspective.
In R. Vasta (Ed.), Annals of child development: A research
annual (Vol. 10, pp. 167–208). London: Jessica
Kingsley Publishers.
Cummings, E. M., Davies, P. T., & Simpson, K. S. (1994).
Marital conflict, gender, and children’s appraisals and
coping efficacy as mediators of child adjustment. Journal
of Family Psychology, 8, 141–149.
Cummings, E. M., & Smith, D. (1993). The impact of anger
between adults on siblings emotional and social behavior.
Journal of Child Psychology and Psychiatry, 34, 1425–1433.
Davies, P. T., & Cummings, E. M. (1994). Marital conflict
and child adjustment: An emotional security hypothesis.
Psychological Bulletin, 116, 387–411.
DeWolff, M. S., & van Ijzendoorn, M. H. (1997). Sensitivity
and attachment: A meta-analysis on parental antecedents
of infant attachment. Child Development, 68, 571–591.
Dunn, J. (1993). Young children’s close relationships: Beyond attachment.
Newbury Park, CA: Sage.
Egeland, B., & Farber, E. A. (1984). Infant-mother attachment:
Factors related to its development and changes
over time. Child Development, 55, 753–771.
Egeland, B., & Sroufe, L. A. (1981). Developmental sequelae
of maltreatment in infancy. New Directions in Child Development,
11, 77–92.
Elicker, J., Englund, M., & Sroufe, L. A. (1992). Predicting
peer competence and peer relationships in childhood
from early parent-child relationships. In R. Parke & G.
Ladd (Eds.), Family-peer relationships: Modes of linkage
(pp. 77–106). Hillsdale, NJ: Erlbaum.
Fauber, R., Forehand, R., Thomas, A. M., & Wierson, M.
(1990). A mediational model of the impact of marital
conflict on adolescent adjustment in intact and divorced
families: The role of disrupted parenting. Child Development,
61, 1112–1123.
Fendrich, M., Weissman, M. M., & Warner, V. (1991). Longitudinal
assessment of major depression and anxiety disorders
in children. Journal of the American Academy of
Child and Adolescent Psychiatry, 30, 38–42.
Fiese, B. H., Wilder, J., & Bickham, N. L. (in press). Family
context in developmental psychopathology. In M. Lewis,
A. Sameroff, & S. Miller (Eds.), Handbook of developmental
psychopathology (2nd ed.). New York: Plenum Press.
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon,
G., Target, M., & Gerber, A. (1996). The relation of
attachment status, psychiatric classification, and response
to psychotherapy. Journal of Consulting and Clinical
Psychology, 64, 22–31.
Fonagy, P., Steele, M., & Steele, H. (1991). Maternal representations
of attachment during pregnancy predict the
organization of infant-mother attachment at one year of
age. Child Development, 62, 880–893.
Fury, G., Carlson, E. A., & Sroufe, L. A. (1997). Children’s
representations of attachment relationships in family
drawings. Child Development, 68, 1154–1164.
George, C., Kaplan, N., & Main, M. (1985). The attachment interview
for adults. Unpublished manuscript, Department
of Psychology, University of California, Berkeley.
Grych, J. H., & Fincham, F. D. (1990). Marital conflict and
Lewis, Feiring, and Rosenthal 719
children’s adjustment: A cognitive-contextual framework.
Psychological Bulletin, 108, 267–290.
Grych, J. H., & Fincham, F. D. (1993). Children’s appraisals
of marital conflict: Initial investigation of the cognitive
contextual framework. Child Development, 64, 215–230.
Guidbaldi, J. (1988). Differences in children’s divorce adjustment
across grade level and gender: A report from the
N.A.S.P.–Kent State Nationwide Project. In S. A. Wolchick
& P. Karoly (Eds.), Children of divorce: Empirical perspectives
on adjustment (pp. 185–231). New York: Gardner.
Hamilton, D. L. (1981). Illusory correlation as a basis for stereotyping.
In D. L. Hamilton (Ed.), Cognitive processes in
stereotyping and intergroup behavior (pp. 115–144). Hillsdale,
NJ: Erlbaum.
Harold, G. T., & Conger, R. D. (1997). Marital conflict and
adolescent distress: The role of adolescent awareness.
Child Development, 68, 333–350.
Hastie, R. (1981). Schematic principles in human memory.
In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social
cognition: The Ontario Symposium (Vol. 1, pp. 39–88).
Hillsdale, NJ: Erlbaum.
Hazan, C., & Shaver, P. (1987). Conceptualizing romantic
love as an attachment process. Journal of Personality and
Social Psychology, 52, 511–524.
Henry, B., Moffitt, T. E., Caspi, A., Langley, J., & Silva, P. A.
(1994). On the “Remembrance of things past”: A longitudinal
evaluation of the retrospective method. Psychological
Assessment, 6, 92–101.
Hetherington, E. M., & Clingempeel, W. G. (1992). Coping
with marital transitions: A family systems perspective.
Monographs of the Society for Research in Child Development,
57(2–3, Serial No. 227).
Hetherington, E. M., Cox, M., & Cox, R. (1982). Effects of divorce
on parents and children. In M. E. Lamb (Ed.), Nontraditional
families (pp. 233–288). Hillsdale, NJ: Erlbaum.
Howes, P., & Markman, H. J. (1989). Marital quality and
child functioning: A longitudinal investigation. Child
Development, 60, 1044–1051.
Hudson, J. A. (1990). The emergence of autobiographical
memory in mother-child conversation. In R. Fivush & J.
Hudson (Eds.), Knowing and Remembering in Young Children
(pp. 166–196). Cambridge, U.K.: Cambridge University
Press.
Hudson, J. A., Fivush, R., & Kuebli, J. (1992). Scripts and
episodes: The development of event memory: Special
Issue: Memory in everyday settings. Applied Cognitive
Psychology, 6, 483–505.
James, W. (1975). Pragmatism (F. Bowers, Ed.). Cambridge,
MA: Harvard University Press. (Original work published
1895).
Kobak, R., & Sceery, A. (1988). Attachment in late adolescence:
Working models, affect regulation, and representations
of self and others. Child Development, 59, 135–
146.
Kobak, R. R., Cole, H. E., Ferenz-Gillies, R., Fleming, W. S.,
& Gamble, W. (1993). Attachment and emotion regulation
during mother-teen problem solving: A control theory
analysis. Child Development, 64, 231–245.
Kobak, R. R., Sudler, N., & Gamble, W. (1991). Attachment
and depressive symptoms during adolescence: A developmental
pathways analysis. Development and Psychopathology,
3, 461–474.
Lamb, M. E., Thompson, R., Gardner, W., & Charnov, E.
(1985). Infant-mother attachment: The origins and developmental
significance of individual differences in strange situation
behavior. Hillsdale, NJ: Erlbaum.
Lerner, R. M. (1986). Concepts and theories of human development
(2nd ed.). New York: Random House.
Lewis, M. (1984). Social influences on development: An
overview. In M. Lewis (Ed.), Beyond the dyad (pp. 1–12 ).
New York: Plenum Press.
Lewis, M. (1997). Altering fate: Why the past does not predict
the future. New York: The Guilford Press.
Lewis, M., & Feiring, C. (1989). Infant, mother, and motherinfant
interaction behavior and subsequent attachment.
Child Development, 60, 831–837.
Lewis, M., & Feiring, C. (1991). Attachment as personal characteristic
or a measure of environment. In J. L. Gewirtz &
W. M. Kurtines (Eds.), Intersections with attachment (pp. 1–
21). Hillsdale, NJ: Erlbaum.
Lewis, M., Feiring, C., McGuffog, C., & Jaskir, J. (1984). Predicting
psychopathology in six-year-olds from early social
relations. Child Development, 55, 123–136.
Lewis, M., & Starr, M. (1979). Developmental continuity. In
J. Osofsky (Ed.), Handbook of infant development (pp. 653–
670). New York: Wiley.
Loftus, E. F. (1981). Mentalmorphosis: Alterations in memory
produced by the mental bonding of new information
to old. In J. B. Long & A. D. Baddeley (Eds.), Attention and
performance IX (pp. 417–433). Hillsdale, NJ: Erlbaum.
Loftus, E. F., Miller, D. G., & Burns, H. J. (1978). Semantic integration
of verbal information into visual memory. Journal
of Experimental Psychology: Human Learning and Memory,
4, 19–31.
Londerville, S., & Main, M. (1981). Security of attachment,
compliance, and maternal training methods in the second
year of life. Developmental Psychology, 17, 298–299.
Main, M. (1990). Cross-cultural studies of attachment organization.
Recent studies, changing methodologies, and
the concept of conditional strategies. Human Development,
33, 48–61.
Main, M., & Goldwyn, R. (1991). Adult Attachment Classification
System. Unpublished manuscript, University of California,
Berkeley.
Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy,
childhood, and adulthood: A move to the level of
representation. In I. Bretherton & E. Waters (Eds.), Growing
points of attachment theory and research. Monographs
of The Society for Research in Child Development, 50
(1–2, Serial No. 209), 66–104.
Matas, L., Arend, R. A., & Sroufe, L. A. (1978). Continuity of
adaptation in the second year: The relationship between
quality of attachment and later competence. Child Development,
49, 547–556.
Oppenheim, D., Emde, R. N., & Warren, S. (1997). Children’s
narrative representations of mothers: Their development
and associations with child and mother adaptation.
Child Development, 68, 127–138.
Pastor, D. L. (1981). The quality of mother-infant attach-
720 Child Development
ment and its relationship to toddlers’ initial sociability
with peers. Developmental Psychology, 17, 326–335.
Pepper, S. C. (1942). World hypotheses. Berkeley: University
of California Press.
Posada, G., Waters, E., Crowell, J. A., & Lay, K. (1995). Is it
easier to use a secure mother as a secure base? Attachment
Q-Sort correlates of the adult attachment interview.
Monographs of the Society for Research in Child Development,
60(2–3, Serial No. 244).
Rosenstein, D. S., & Horowitz, H. A. (1996). Adolescent attachment
and psychopathology. Journal of Consulting and
Clinical Psychology, 64, 244–253.
Ross, M. (1989). Relation of implicit theories to the construction
of personal histories. Psychological Review, 96, 341–357.
Sameroff, A. J., Bartko, W. T., Baldwin, A., Baldwin, C., & Seifer,
R. (1998). Family and social influences on the development
of child competence. In M. Lewis & C. Feiring (Eds.),
Families, risk, and competence. Mahwah, NJ: Erlbaum.
Schneider, W., & Bjorklund, D. F. (1998). Memory. In D.
Kuhn & R. S. Siegler (Eds.), W. Damon (Series Ed.),
Handbook of child psychology: Vol. 2. Cognition, perception,
and language. (5th ed., pp. 467–521). New York: Wiley.
Sroufe, L. A. (1983). Infant-caregiver attachment and patterns
of adaptation in preschool: The roots of maladaptation
and competence. In M. Perlmutter (Ed.), Development
and policy concerning children with special needs: The
Minnesota Symposia on Child Psychology (Vol. 16, pp. 41–
83). Hillsdale, NJ: Erlbaum.
Sroufe, L. A. (1997). Psychopathology as an outcome of development.
Development and Psychopathology, 9, 251–268.
Sroufe, L. A., Egeland, B., & Kreutzer, T. (1990). The fate of
early experience following developmental change: Longitudinal
approaches to individual adaptation in childhood.
Child Development, 61, 1363–1373.
Sroufe, L. A., & Fleeson, J. (1986). Attachment and the construction
of relationships. In W. Hartup & Z. Rubin
(Eds.), Relationships and development (pp. 239–252). Hillsdale,
NJ: Erlbaum.
Sroufe, L. A., Schork, E., Motti, E., Lawroski, N., & LaFreniere,
P. (1984). The role of affect in social competence. In
C. Izard, J. Kagan, & R. Zajonc (Eds.), Emotional cognition
and behavior. New York: Plenum Press.
Suomi, S. J. (1979). Differential development of various social
relationships by Rhesus monkey infants. In M. Lewis & L.
Rosenblum (Eds.), The child and its family: The genesis of behavior
(Vol. 2, pp. 219–244). New York: Plenum Press.
Taylor, S. E., & Crocker, J. (1981). Schematic bases of social
information processing. In E. T. Higgins, C. P. Herman,
& M. P. Zanna (Eds.), Social cognition: The Ontario Symposium
(Vol. 1, pp. 89–134). Hillsdale, NJ: Erlbaum.
Thompson, R. A., & Lamb, M. E. (1983a). Individual differences
in dimensions of socioemotional development in
infancy. In R. Plutchik & H. Kellerman (Eds.), Emotion:
Theory, research, and experience: Vol. 2. Emotions in early development.
New York: Academic Press.
Thompson, R. A., & Lamb, M. E. (1983b). Security and attachment
and stranger sociability in infancy. Developmental
Psychology, 19, 184–191.
U.S. Department of Health and Human Services. (1995). Vital
statistics of the United States.
van Ijzendoorn, M. H. (1995). Adult attachment representations,
parental responsiveness, and infant attachment:
A meta-analysis on the predictive validity of the Adult
Attachment Interview. Psychological Bulletin, 117, 387–
403.
van Ijzendoorn, M. H., Kranenburg, M. J., ZwartWoudstra,
H. A., von Busschbach, A. M., & Lambermon,
M. W. (1991). Parental attachment and children’s socioemotional
development: Some findings on the validity
of the Adult Attachment Interview in the Netherlands.
International Journal of Behavioral Development, 14, 375–
394.
Vaughn, B., Egeland, B., Sroufe, L. A., & Waters, E. (1979).
Individual differences in infant-mother attachment at
twelve and eighteen months: Stability and change in
families under stress. Child Development, 50, 971–975.
Ward, M. J., & Carlson, E. A. (1995). The predictive validity
of Adult Attachment Interview for Adolescent Mothers.
Child Development, 66, 69–79.
Waters, E., & Dean, K. E. (1982). Infant-mother attachment:
Theories, models, recent data and some tasks for comparative
developmental analysis. In L. W. Hoffman & R. J.
Gandelman (Eds.), Parenting: Its causes and consequences
(pp. 19-54). Hillsdale, NJ: Erlbaum.
Waters, E., Posada, G., Crowell, J., & Lay, K. (1993). Is attachment
theory ready to contribute to our understanding
of disruptive behavior problems? Development and
Psychopathology, 54, 215–224.
Waters, E., Treboux, D., Crowell, J., Merrick, S., & Albersheim,
L. (1995, April). From the strange situation to the
adult attachment interview: A 20-year longitudinal study of
attachment security in infancy and early adulthood. Poster
Session presented at the biennial meeting of the Society
for Research in Child Development, Indianapolis, IN.
Waters, E., Wippman, J., & Sroufe, L. A. (1979). Attachment,
positive affect, and competence in the peer group: Two
studies in construct validation. Child Development, 50,
821–829.
Wohlwill, J. F. (1973). The study of behavioral development.
New York: Academic Press.
Yarrow, M. R., Campbell, J. D., & Burton, R. V. (1970). Recollections
of childhood: A study of the retrospective
method. Monographs of The Society for Research in Child
Development, 35(5, Serial No. 138). Chicago: University of
Chicago Press.
Zill, N. (1988). Behavior, achievement, and health problems
among children in step families: Findings from a national
survey of child health. In E. M. Hetherington & J. D.
Arasteh (Eds.), Impact on divorce, single parenting, and stepparenting
on children (pp. 325–368). Hillsdale, NJ: Erlbaum.
Zimiles, H., & Lee, V. E. (1991). Adolescent family structure
and educational progress. Developmental Psychology, 27,
314–320.
Zimmerman, P., Fremmer-Bombik, E., Spangler, G., &
Grossman, K. E. (1997). Attachment in adolescence: A
longitudinal perspective. In W. Koops, J. B. Hoeksema, &
D. C. vandenBoom (Eds.), Development of interaction and
attachment: Traditional and non-traditional approaches (pp.
281–291). Amsterdam: North-Holland.