Insecure Attachment 265attached to one parent might, if reassessed with other caregivers, turn
out to be insecurely attached to the other parent, to a nanny, or to a
day care worker. Thus, while some young children are similarly attached
to all their caregivers, others have a mixture of secure and insecure at-
tachments. For this latter group with mixed attachment types, the current
evidence suggests that subsequent development is particularly affected by
the quality of attachment to the most important caregiver. Which caregiver
is most important? The amount of contact seems relevant since several
studies have shown that when infants spend most of their waking time
in day care, their development is better predicted by the quality of their
attachment to day care staff than by the quality of their attachment to
parents. One important implication is that ‘quality time’ with parents may
not entirely offset the ill effects on infants of prolonged low-quality day
care. Behaviour genetic studies in twins suggest attachment patterns are
predominantly environmentally determined, with no heritable compo-
nent, suggesting that they are indeed formed by parenting behaviour.
Attachment throughout life
While the SSP tests the attachment security of children aged 12–18
months, newer assessments are emerging for older children. These newer
assessments also generate ABCD categories on the basis of the child’s
response to reunion with a caregiver.
Thus, for example for children aged 4-8, doll-play tasks can be used. The
child is asked to name dolls to match who is at home (for example, me, my
mother and my brother) then a number of lively scenarios are commenced
by an interviewer that are designed to trigger attachment behaviour (for
example, an accident leading to a cut knee). At this juncture the child is
asked to enact what happens next. If the ‘me’ person as enacted by the
child gets well looked after by the mother, a secure designation is likely,
whereas if the parent is indifferent or the child seeks no solace, one of the
insecure categories is more likely to be assigned. Insecure patterns elicited
by such ‘story-stems’ are predictive of psychopathology. In adolescence, a
modification of the Adult Attachment Interview (see below) called the
Child Attachment Interview has been devised for 8–15-year-olds, and
again predicts psychopathology. Moreover, insecurity predicts symptoms
over and above what is associated with poor parenting, suggesting that
there is indeed value to the notion that an internal working model
organises the child responses.
A child’s attachment classification is moderately stable in childhood, age
5 year stability being of the order of 0.4 or 0.5. One explanation for this
modest stability is that a child’s internal working model of relationships
is largely determined early in childhood and is then resistant to change.
It should be borne in mind that assessment methods change as noted
above, so some of the discontinuity may be due to different measurement
methods. An alternative explanation, for which this increasing evidence,