Coming soon: AboutKidsHealth is getting a new look! Learn more Watch a video tour

print article
For optimal print results, please use Internet Explorer, Chrome or Safari.

Attachment Part Three: Attachment across the life span

Ainsworth’s Strange Situation Paradigm revealed that most infants respond in one of a few predictable ways to separations from caregivers; a major discovery for attachment researchers. But what is the wider significance of this finding?

In Parts One and Two of our series on attachment, we saw how different patterns of attachment arise during infancy. In part three, Dr. Goldberg explores the development of attachment patterns beyond infancy and across the lifespan.

By Susan Goldberg, PhD

For Bowlby, the founder of attachment theory, attachment involved much more than an infant’s response to separation. Infant separation and reunion behaviours were just a preview of things to come; one indication of an underlying pattern of expectations about the self, others, and the world which would continue to evolve and persist throughout adulthood.

Bowlby’s ideas were revolutionary and controversial, generating a flurry of questions; what were the long-term implications of early attachment security? What, if anything, do our earliest attachment relationships tell us about the kind of parents we will become? And can our attachment history with our parents affect the security and well-being of our own children?

Continuing attachment research is providing answers to these fundamental questions. There is mounting evidence that security of attachment can affect many aspects of a person’s well-being in profound and enduring ways, and that these outcomes can in turn be conveyed to future generations.

Beyond infancy

In 1974, researchers at the Institute of Child Development at the University of Minnesota began a longitudinal study that would span three decades and prove to be a landmark in attachment research. From an original sample of 267 expectant mothers came a multitude of findings pointing to the critical importance of early relationships. What was once speculation now had empirical support.

Infant-caregiver pairs were assessed at one year using the Strange Situation Paradigm. A group of children were re-assessed between the ages of four and five. Months of intensive observation revealed that the majority of children classified as secure at one year of age  scored higher on measures of self-esteem, responded more positively to other children, and had better social skills than children in insecure attachmant relationships.

New measures of attachment

In order to assess the stability of attachment over time, researchers developed new measures that could evaluate and classify attachment patterns in older children and adults:

  • Children up to seven years of age are assessed using observational procedures. Following separations between children and attachment figures, affective expressions, conversational exchanges, and subtle body language are observed and evaluated during reunions.
  • The Adult Attachment Interview (AAI) is the adult and adolescent equivalent of the Strange Situation Paradigm. The AAI has recently been adapted for use with children as young as 7 years old. Participants are asked about early experiences with attachment figures. Attachment security is reflected by the degree to which respondents coherently discuss childhood attachment experiences. Both the answers and the manner in which they are expressed are evaluated.

With the AAI and various observational measures, classifications schemes that correspond to the four infant-caregiver attachment patterns have been developed.

Preschool and school aged attachment

The Minnesota studies and many that followed showed promising outcomes for most children who were in secure attachment relationships with their primary caregiver during infancy. Children who were in secure attachment relationships at one year of age tend to be popular with peers, resilient, resourceful, and cooperative in preschool. By age six, they are more compliant, responsive, cooperative, self-reliant, and empathic than those who were in insecure attachment relationships in infancy. In general, secure attachment appears to act as a protective factor against emotional and behavioural problems in childhood and adolescence.

Infants who display avoidant attachment in the Strange Situation Paradigm at one year are at greater risk for a defended or avoidant attachment patterns during preschool; especially in the presence of other risk factors. Children who were in avoidant attachment relationships during infancy are vulnerable to becoming emotionally insulated, hostile, or anti-social; engaging in activities such as stealing, lying, or cheating. They are more likely to provoke adults and peers into rejecting them and are more likely to victimize others than those who were in secure relationships during infancy.

Children who display resistant attachment at one year of age tend to be dependent during the preschool years. Dependent children spend a disproportionate amount of time seeking attention from adults. They may be easily frustrated or passive and helpless. Like infants in resistant attachment relarionships, dependent children are preoccupied with the caregiver at the expense of other activities.

Children who show disorganized patterns of attachment in the Strange Situation Paradigm at one year of age have the greatest risk of aggressive behaviour, conduct disorder, and dissociative behaviours later in life. Dissociative behaviours involve a breakdown in a person’s perception of his or her surroundings, memory, identity, or consciousness.

By age six, disorganized attachment has become organized but into a pattern where the child, rather than the caregiver takes responsibility for control. These children control, coerce, or dominate their caregiver either by humiliating and rejecting her or by being attentive and protective. The controlling-caregiving child entertains or comforts the caregiver who usually shows limited emotion in response to the child’s often exaggerated positive emotions. The controlling-punitive-child behaves in a hostile manner toward the caregiver who usually complies with the child’s requests.

A final childhood attachment classification, insecure-other, includes children who show insecurity but do not fit into the other categories. These children tend to show a mixture of avoidant and dependent strategies.

Adolescent and adult attachment

Autonomous attachment in adolescence and adulthood corresponds to the infant-caregiver secure attachment pattern in infancy. These individuals can recognize both the limitations and positive qualities of childhood attachment figures. Individuals classified as autonomous value relationships, are often forgiving of less than optimal caregiving, are at peace with imperfections in themselves and others, and are coherent and objective in describing their childhood experiences.

The classification of preoccupied attachment corresponds to insecure resistant attachment in infancy. Individuals classified as preoccupied get very entangled in the details of childhood but find it difficult to provide a clear overview of their past. They tend to be preoccupied with early experiences, often appearing overly concerned with trying to please their parents. Adults and adolescents classified as preoccupied are not at peace with imperfections in themselves or others.

Dismissing patterns of attachment corresponds most closely with insecure-avoidant attachment in infancy. In responding to the AAI, these individuals tend to say little about childhood attachment experiences often providing only short, minimally informative answers and insisting upon an inability to recall while often presenting an overly positive account of early childhood experiences. They also minimize the impact of key relationships from their past.

Unresolved attachment, which corresponds most closely to disorganized attachment in infancy, is characterized by a lack of resolution of mourning of a significant loss or trauma. Individuals who are classified as unresolved in their attachment style may show confusion surrounding a death or trauma, confusion about the permanency of a death, or a sense of being possessed by the deceased person or abuser.

Stability and change in attachment patterns

Attachment theory assumes that attachment patterns can endure across the lifespan. The adoption of a particular attachment strategy in infancy generally predicts a specific developmental trajectory that carries on throughout a person’s life. Definitive evidence confirming the stability of attachment would require intensive life-time follow-up procedures. At present, there are less comprehensive, often mixed, but nevertheless valuable findings.

The Family Lifestyles Project is an ongoing longitudinal study based at the University of California. Following children from infancy to 19 years of age, attachment during infancy was found to be a significant predictor of attachment security during adolescence.

In contrast to these findings, a sample of 57 high risk young adults showed no significant continuity of attachment style between infancy and adulthood. The sample consisted of a subset of participants from the Minnesota Mother-Child Project, an ongoing longitudinal study of children at risk for poor developmental outcomes. Continuity of attachment depended on specific conditions of adversity. Infants who suffered from maltreatment and had insecure attachment relationships with their primary caregiver were more likely to remain insecure. Infants in insecure attachment relationships who became secure as young adults, known as earned secure, were more likely to have better family functioning at age 13. And finally, infants who received secure attachment classifications in the Strange Situation Paradime and received insecure classifications in later years, were more likely to have depressed mothers.

Another 20-year longitudinal study, based in Minneapolis, looked specifically at the impact of negative life events on the stability of attachment. Of the participants who had experienced significant loss or stress, 44% changed attachment classifications from infancy to early adulthood, in contrast to the only 22% of individuals who did not experience negative life events.

Intimate relationships

We've seen how attachment security can continue to influence core aspects of a person’s life well into adulthood. How do these effects in turn affect the quality of significant relationships in people’s lives? Although there is little longitudinal research linking infant-caregiver attachment and adult intimate relationships, there is some evidence to suggest that marital satisfaction can be related to a person’s attachment classification on the AAI. Couples with one insecure partner experience more conflict than couples with both autonomous partners; two insecure partners result in the greatest degree of conflict. Among engaged and married couples, insecure women report more verbal and physical aggression and threats of abandonment from partners compared to couples in which the women are autonomous.

The transmission of attachment to our children

The quality of our intimate relationships can play a role in our ability to be effective parents for our children. As well, there is evidence to suggest that a person’s parenting style can mimic the style to which they were exposed as infants. Thus, adults with autonomour attachment tend to consistently respond in comforting ways to their infants’ distress. Adults who had dismissing and preoccupied attachment relationships with their parents are more likely to be rejecting or inconsistent respectively as caregivers, thus perpetuating the likelihood of insecure attachment in their infants.

Diane Benoit from the Hospital for Sick Children and Kevin Parker from Kingston General Hospital were interested in the degree to which attachment classification extended across generations. They found that mothers’ attachment classifications, collected during pregnancy successfully predicted their infants’ classification at twelve months in 82% of cases. Looking across three generations, 65% of the grandmother-mother-infant triads had corresponding attachment classifications in all three generations. This finding extends the already far-reaching implications of early attachment security.

In line with Bowlby’s vision of attachment as a lifespan construct, there is mounting evidence that the long term effects of early attachment relationships can be far-reaching. Our attachment history can affect our emotional well-being which in turn can influence our friendships, our choice of significant others, and the quality of our romantic relationships. The strength of core relationships affects the kind of caregivers we become and ultimately our own children’s sense of security and emotional well-being.

Continuing research on stability and change in attachment patterns will shed more light on this critical area of human development.

In Part 4 of our series we will explore the particular child and caregiver characteristics that contribute to attachment security during development.

Diane Benoit, MD, FRCPC

Sheri Madigan, PhD, C.Psych (supervised practice)


Benoit D, Parker KCH. Stability and transmission of attachment across three generations. Child Development. 1994;65:1444-1456.

Waters E, Hamilton CE,Weinfield NS. The stability of attachment security from infancy to adolescence and early adulthood: General introduction. Child Development. 2000;71(3),678-683.