Search
Program Calendar
Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
Search Tips
Virtual Exhibit Hall
Personal Schedule
Sign In
Main et al. (1985)’s discovery that parents’ attachment representations are associated with parent-child attachment has intrigued researchers across the globe. Research has focused on caregiver sensitivity as the mechanism through which attachment may be transmitted to children, but meta-analytic evidence showed that sensitive responsiveness explained only 25% of the association (Van IJzendoorn, 1995; Verhage et al., 2016). Solving this so-called ‘transmission gap’ has been a major focus, but exploration of other potential mediating variables has not yet resolved this issue. A possible explanation for this conundrum is that the mediating pathway may differ between parent-child dyads based on their background characteristics. A high-risk background (e.g., teenage parents, maternal mental health problems, child health problems) has recently been shown to moderate the association between parents’ attachment representations and parent-child attachment (Verhage et al., 2018): within high-risk dyads, transmission of autonomous to secure attachment was less likely than in low-risk dyads, but transmission of non-autonomous to insecure was equally likely in both groups. The impact of moderators on the mediational pathways of attachment transmission through caregiver sensitivity has not yet been examined. The current study tested the hypothesis that the mediation model of attachment transmission through caregiver sensitivity differed between high-risk and low-risk dyads.
The study was performed in a large consortium for the synthesis of studies on attachment transmission, comprising 58 studies (total N=4,396 parent-child dyads) which collated data on parental attachment representations and parent-child attachment. These preliminary analyses are based on a subset of 2,287 parent-child dyads (from 32 studies) with additional data on parental sensitivity. Forced dichotomized classifications for attachment representations and parent-child attachment were used; AQS scores (N=692) were recoded (insecure < .40, secure ≥ .40, following Lehman et al., 1992).
Multilevel mediation modeling confirmed partial mediation by sensitivity in the association between attachment representation and parent-child attachment. The indirect effect was b=0.13, 95% CI [ 0.08, 0.17], p<.001, accounting for 16% of the total effect (b=0.83, 95% CI [0.64, 1.01], p<.001). To assess the moderating effect of risk background, a multilevel moderated mediation model with intercept constrained to be equal across groups was fitted to the data (Figure 1). The likelihood ratio test comparing this model with a model with all parameters constrained showed better fit for the model with distinct parameters for high-risk (N=971) and low-risk (N=1,316) dyads, LR χ2 (9) = 50.47, p<.0001. Wald tests to assess equality of specific paths between high-risk and low-risk groups indicated that the path between attachment representations and sensitivity was equal (χ² = 2.23, p=.14), the path between sensitivity and parent-child attachment was not equal (χ² = 7.31, p=.002), and the direct path between attachment representations and parent-child attachment was not equal (χ² = 11.09, p<.001).
Findings indicate that dyadic background characteristics not only impact intergenerational transmission of attachment, but also the role of sensitive behavior in individual differences in attachment.
The Collaboration on Attachment Transmission Synthesis
Marije L. Verhage, Vrije Universiteit Amsterdam
Presenting Author
Pasco Fearon, Research Department of Clinical, Educational and Health Psychology, University College London
Non-Presenting Author
Carlo Schuengel, Vrije Universiteit Amsterdam
Non-Presenting Author
Marinus H van IJzendoorn, Erasmus University Rotterdam
Non-Presenting Author
Marian J Bakermans-Kranenburg, Vrije Universiteit Amsterdam
Non-Presenting Author
Sheri Madigan, University of Calgary and the Alberta Children’s Hospital Research Institute, Calgary
Non-Presenting Author
Glenn Roisman, University of Minnesota
Non-Presenting Author
Mirjam Oosterman, Vrije Universiteit Amsterdam
Non-Presenting Author