Attachment

Origins

Development

Intergenerational transmission

What is it?

  • Intimate bond between baby and primary caregivers
  • Behaviour serving to maintain proximity to a selective caregiver(s) in times of stress --> triggered by cues of danger, brings about proximity and feeling of safety
  • Theorised evolutionary basis, develops early in infancy, most clearly evident at 7-9m by proximity seeking and stranger anxiety
  • Ainsworth (1967, 1973) proposed that differences in the security of infant–mother attachment have sig. long-term implications for later intimate relationships, self-understanding, and even risk for psychopathology.

Attachment patterns:

  • Secure Attachment: seek proximity, communicate need for comfort, contact is effective
  • Avoidant Attachment: avoids contact, minimizes expressions of need for contact
  • Resistant Attachment: intense expression of distress, angry upon contact, contact not effective
  • Disorganized Attachment: contradictory, fragmented, disoriented or fearful behaviour upon contact e.g. wondering around the room, other stereotypical behaviours (pulling on hair)
  • Disinhibited Attachment: extreme social disinhibition, lack of stranger caution, approach and receive comfort from strangers

Child usually has more than one caregiver/attachment figure = can have more than one attachment pattern, e.g. is it more important to have secure attachment with mother than father
Westernised phenomenon – very different family dynamics in the East e.g. more children per family, not typical mother father dynamic


  • Security associated with sensitivity, defined by awareness of infant attachment cues, accurate interpretation of infant cues, responsive to cues, appropriate response
  • Insecurity associated with insensitive care --> negative/rejecting, interfering/intrusive, inconsistent availability
  • Disorganization associated with frightening, frightened parenting, maltreatment
  • Disinhibited Attachment associated with institutional care, extreme neglect

In secure attachment, potential threat = avoidance of threat and proximity seeking towards caregiver
If source of danger is coming from caregiver then they want proximity from their caregiver yet as they get closer avoid because the caregiver is the source of the fear = causes a behavioural paradox --> frightened of the person you should go to when you’re frightened

  • Bokhorst et al. (2003) 157 pairs MZ and DZ twins.
    52% of the variance in attachment security was explained by shared environment (vs. nonsecure) and 48% of the variance was explained by unique environmental factors and measurement error.
    Little evidence of genetic or shared E influence on disorganisation.
    Genetic factors explained 77% of the variance in temperamental reactivity, and unique environmental factors and measurement error explained 23% --> not associated with attachment concordance.
  • Luijk et al (2011) 2 large samples: Generation R study (N = 663), NICHD Study of Early Child Care (N = 478-522), 12 different genes (including DRD4, 5HTTLPR) --> No consistent genetic effects or gene x parenting interactions across either sample except small effect of COMT gene
    = attachment and parenting are strongly shared environmental variables with little evidence of genetic influence (= attachment provides a psychological and environmental account of intergenerational transmission)

Biological processes:

  • Viewed as evolved behavioural system- infants seek proximity when distressed as it promotes survival to reproductive success; insecure attachment also adaptive- encourage search for alternative attachment figures if adults consistently unresponsive/resistant
  • Belsky et al. (1991) applied to life history theory- quality of parental care sensitises to the supportiveness/aversiveness of E, and this easy experience --> security of attachment and reproductive success adaptations (e.g. timing of pubertal maturation, onset of sexual activity, preferences in pair bonding, eventual parental investment in own offspring)
    = children in high stress/insecure families --> develop reproductive strategies that are low-investment and opportunistic. Secure, low stress families opposite pattern.
  • Thompson (2015): early experiences of chronic stress (e.g. maternal depression, emotional inaccessibility & abusive events) alter the neurocircuitry of stress reactivity = develop dysregulated patterns of stress responding that determine self-regulation, threat vigilance and blunt attentional focus
  • Social support buffers against the effect of stress = can explain insecure/secure attachment --> secure enables better emotion regulation and cognitive functioning

Externalising problems:

  • Fearon et al. (2010) meta analysis, 69 samples --> sig. association between insecure attachment and externalising problems. Larger effects for boys, clinical samples, observation-based outcome assessments, and attachment assessments other than the SS. Disorganised children at risk, with weaker effects for avoidant and resistant attachment styles
  • Groh et al. (2012) meta analysis of 42 samples looking at association between attachment and internalising symptoms --> only small association with insecurity but was sig. Not moderated by assessment age. Avoidance, but not resistance or disorganisation sig. associated with internalising symptoms. Insecurity and disorganisation more strongly associated with externalizing than internalizing symptoms.
  • Bohlin et al. (2012) disorganised attachment at age 5 predicted CU traits at age 7 --> linked to CP
  • Dadds et al. (2011) children w/ high levels CU traits made less eye contact with both mother and farther, and also lower levels of physical and verbal affection in a free play situation = attribute to attachment issues

Personality:

  • Minnesota Study of Risk and Adaptation from Birth to Adulthood (Sroufe et al. 2005), largest study of attachment and development. Children followed from infancy-34, with SS conducted at 12 & 18m and personality characteristics regularly assessed through observations, observer ratings and self-reports.
    -Sig. association b/ early attachment and personality through childhood and adolescence --> emotional health, self-esteem, agency and self-confidence, positive affect, ego resiliency and social competence in interactions with peers, teachers, romantic partners and others.
    -As children matured, importance of early attachment was determined by subsequent developmental influences e.g. as time progressed between Strange Situation assessments and later personality outcomes, the effects of early security were more likely to be indirect—mediated and/or moderated by subsequent relational influences

Association between security of attachment and later measures of cognitive performance and IQ

  • more carefully designed mediational studies show that this is because of differences in parental quality of assistance, peer relationships, and children’s cooperativeness at school—mediators that are fully consistent with attachment theory (Drake et al. 2014)

IWM

Attachment sensitivity--> Internal Working Models --> self confidence, empathy/moral development, social attributions, emotion-regulations, prosociality --> behaviour problems

Internal working models: key concept in attachment theory

  • According to Bowlby- attachment security influences psychological growth through children's developing mental representations (IWMs) of the social world, based on the accessibility and responsiveness of their caregivers
  • These expectations --> broader representations of their attachment figures, interpretations of their relational experiences, guidelines about how to interact with others, and even beliefs about themselves as relational partners
    --> reconstruct experience of new relationships in ways consistent with past experiences, and expectations arising from secure/insecure attachments = choose new partners and behave with them in ways consistent with expectations created from earlier attachments

Different in how they function:

  • 1) IWMs govern information processing – securely attached more likely to process a broad range of +ve and -ve info related to attachment concerns in a +vely biased manner, but insecure attached more likely to defensively exclude info that leads to psychological pain (e.g. forgets abandonment in childhood) but if other info this will be processed in negatively biased fashion (Dykas & Cassidy, 2011)

  • 2) focus on content of IWMs (Thompson, 2006, 2010)- secure likely to have more constructive representations of other people, more +ve expectations for social interaction, greater social and emotional understanding, more +ve self-concept, and more advanced conscience development com- pared to insecurely attached individuals.


    Means IWMs are shaped by secondary representations, including conversation with adult caregivers, others thoughts, feelings and motivations rather than just the child’s direct relational experience


  • not necessarily mutually exclusive, just have different implications for how IWMs influence social behaviour

Emotion regulation:

  • Thompson (2015) children in secure relationships are stronger in emotion regulation than children in insecure relationships (parents more dismissive, punitive, or critical of the children’s emotional expressions)
  • Gilliom et al. (2002) reported that boys who were securely attached at age 1.5 were observed to use more constructive anger management strategies at age 3.5. Securely attached more likely to use distraction, ask questions about the frustration task, and wait quietly than were insecurely attached boys.
  • Contreras et al. (2000) security in middle childhood sig. associated with children’s constructive coping with stress, and that the measure of coping mediated the association between attachment and children’s peer competence.
    = association between secure attachment and emotion regulation is apparent from infancy to adolescence
  • Securely attached better at enacting behavioural strategies e.g. proximity seeking = likely to result in better emotion management – Leerkes & Wong (2012)
  • Mothers in secure relationships more likely to perceive/interpret more sensitively their children’s emotions = securely attached children have greater depth in their emotion understanding, including their appreciation of effective emotion regulation strategies (Waters & Thompson, 2014)
  • Secure children are more proficient at identifying emotions in others and, in some cases, empathizing with them (Murphy & Laible, 2013) --> link to CPs

Social cognition:

  • Raikes & Thompson (2008): children deemed resistantly attached at 36 months were more likely to make negative motivational attributions to peers as first graders than were secure children. Securely attached at 24 and 36 months more likely to identify socially competent and relevant solutions to social problem-solving tasks than insecure children.
  • controlled for influence of parenting (including maternal sensitivity and depressive symptoms) at multiple assessments to make sure outcomes were a result of early security rather than continuity in parenting

BUT, consistency in parent-child relationship:

  • Securely attached children showed greater enthusiasm, compliance, and positive affect—and less frustration and aggression—during shared tasks with their mothers during the second year (Frankel & Bates, 1990)
  • But longer-term associations between infant security and parent–child interaction at ages 3 (Youngblade & Belsky, 1992) and 5 (Van IJzendoorn 1987) were inconsistent. =depends on consistency in quality of interaction
  • Consistency is mediated by intervening events e.g. family stress, significant changes in family circumstances (such as parental separation or divorce), or other conditions affecting relational harmony (Thompson, 2006)
  • IWMs directly related to child’s ability to create and maintain successful close relationships, establish positive self-image and constructive social representations of people and relationships
  • BUT, concept doesn’t have a rigorously defined theoretical construct (uncertainty about its defining features, functioning and measurement) --> question of whether IWMs are just used for almost anything which a secure attachment is found to be associated with (Belsky & Cassidy, 1994)

Measuring attachment in adults: Adult Attachment interview (AAI)

  • accounts of childhood attachment experiences and the effect of those experiences on present functioning--> Qs about quality/memories of childhood relationships with caregiver, experiences of separation, illness, injury, punishment, loss and abuse
  • Secure-autonomous --> moderate-high coherence; consistency between descriptions/evaluations of attachment experiences and their effects. Avows missing, needing and depending on others.
  • Insecure-dismissing --> Low coherence; high idealization or derogation of parents, claims not to remember details of childhood. Self-described as strong, independent/normal. Little or no articulation of need/dependence
  • Insecure-preoccupied --> Low coherence; passive or angry preoccupation with attachment experiences
  • Unresolved/disorganized --> During discussion of loss/abuse, speaker shows striking lapses in monitoring of reasoning or discourse e.g. belief that a dead person is still alive, disorientation in respect to time.
  • Cannot classify --> scores point to contradictory insecure classifications (e.g. both strong idealization and strong angry preoccupation), or all state-of-mind scores may be low. Also coded if speaker attempts to frighten listener, refuses to speak or seems to fit equally well to a secure and insecure classification

  • Fonagy et al. (1991) --> predictions from attachment class before birth to child's behaviour
    -79% secure mums had secure children, 73% insecure mums had insecure children
  • Intergenerational transmission assumes caregiver's mental representation of attachment --> relationship with child (affects quality of care) --> impacts infants IWM--> mental representation of attachment as adult
  • Meta-analytic evidence from van Ijzendoorn (1995) suggests IT is strong and replicable, but explaining it through maternal sensitivity leaves a "transmission gap" --> maternal sensitivity only explains 25% variance = 75% unaccounted for
  • decades of research on this gap, can be explained by substantive factors (see below branch) and methodological ones
    e.g. longer time intervals between assessment of sensitivity and attachment = smaller effect sizes, training status of coders thought to impact effect size --> larger measurement error with coders without official training (attachment measures typically require intensive training)

Potential moderators of IT? (substantive factors)

  • Verhage et al. (2016) analyses on 95 samples (N=4,819)
  • confirmed IT of attachment - larger effect sizes for secure-autonomous transmission (r=.31) than for unresolved transmission (r=.21) --> sig. smaller effects than for van Ijzendoorn study
  • Effect sizes moderated by risk status of the sample, biological relatedness of child-caregiver dyads, and age of the children (not gender of parent)
  • Multivariate moderator analyses showed that unpublished and more recent studies had smaller effect sizes than published and older studies --> publication bias?? Its reported that 65% of studies finding null results may never even be written up (Franco et al. 2014)
  • Path analyses showed that the transmission could not be fully explained by caregiver sensitivity, with more recent studies narrowing but not bridging the “transmission gap”.
  • Future research: look at more contextual factors (e.g. family functioning, couple relationship, support), neural mechanisms behind attachment and parenting, and effects of differential susceptibility to child-caregiver interaction for children with temperamental characteristics (known to have bidirectional parenting relationship) on IT of attachment

1) risk status of sample:

  • children of depressed mothers (Atkinson et al., 2000) and with adolescent mothers (Cyr et al. 1997) more likely to develop insecure attachment
    = certain children are at higher risk of attachment issues --> might moderate strength of IT

2) Biological vs. non-biological caregivers:

  • van den Dries et al. (2009) meta-analysis compared attachment in adopted children, foster children, and non-adopted children --> rates of insecure and disorganized attachment of foster and adopted children are both higher than these rates in non- adopted children.
  • Unlikely attribute to genetics because not really associated with attachment status, probably to do with shorter history (i.e. not from birth) or that these children often have more -ve experiences

3) Gender of parent:

  • association between sensitivity and attachment weaker in fathers than mothers (De Wolff & Van IJzendoorn, 1997), and the effect size of this association appeared unchanged for fathers over the last 3 decades (Lucassen et al., 2011)
    = effect size of IT of attachment may be smaller for fathers than for mothers

4) Child's age:

  • assumed that quality of attachment experiences is rooted in the history of interactions between the two partners in the relationship = if attachment representations remain a stable influence on those interactions, the strength of IT should increase with age of the child.
  • BUT Van IJzendoorn’s meta-analysis showed smaller effect sizes in several analyses for studies conducted with older children (but small samples)