A Meta-Analysis of the First 20,000 Strange Situation Procedures

• In Event: Attaining greater insight into determinants of parent-child attachment: one (conceptual) model does not fit all

Abstract

Background
For over 50 years, the most relied upon measure to examine individual differences in attachment and exploration behaviors has been the Strange Situation Procedure (SSP). The SSP has now been widely used in both low and high-risk contexts (e.g., socio-economic risk, medical risk, child maltreatment). Theoretically, it has been purported that disorganized attachment would be more likely to develop in high-risk environments, as caregivers may have difficulty perceiving or attending to the infant’s signals and cues for contact and need due to their past or ongoing burdens. In the current study, we take stock of the 287 SSP studies conducted to date by examining the distribution of the four attachment classifications derived in the SSP: secure, avoidant, resistant, and disorganized. A series of moderator analyses will also be conducted to determine if and when distributions of attachment are different across various risk indicators.

Methods:
A science librarian conducted searches in PsycINFO, MEDLINE, EMBASE, Web of Science, and Dissertation Abstracts International for published and unpublished studies from 1967 to February 14, 2020. Database-specific headings and text word fields were searched for concepts of “strange situation” and “attachment”, with truncation symbols used to capture variant endings and spellings (e.g., infant*). Over 15,000 non-duplicate abstracts were identified across the various databases, and 287 studies (and 21,439 dyads) met full inclusion criteria: (1) attachment was assessed in the Strange Situation Paradigm; (2) the study reported the distribution for 4-way attachment (avoidant, secure, resistant, and disorganized attachment), and (3) the SSP was conducted with fathers or mothers. The meta-analysis of the distribution of attachment was conducted using the stata package GLLAMM, which estimates generalized linear mixed models with random effects. Moderator analyses were conducted one by one in separate generalized multinomial random intercept models, in which secure attachment was the reference group.

Results
The overall prevalence of the four attachment categories was estimated as 13.7% Avoidant (95% CI [12.6%, 14.7%]), 53.7% Secure (95% CI [51.6%, 55.7%]), 10.4% Resistant (95% CI [9.4%, 11.4%]) and 22.3% Disorganized (95% CI [20.3%, 24.4%]). Low socio-economic status was associated with higher rates of both avoidant and disorganized attachment, although this association was stronger for disorganized attachment (see Figure 1). Teenage parenthood and child maltreatment were associated with higher rates of avoidant and disorganized attachment. Medical risk was associated with higher rates of disorganized attachment. No risks were associated with raised rates of Resistant attachment.

Conclusions.
The distribution of the SSP classifications in this comprehensive meta-analysis of 287 studies with over 20,000 dyads suggests that secure (53.7%) and disorganized (22.3%) attachment are the two most common classifications. The current study suggests that higher rates of both avoidant and disorganized attachment are expected in contexts of risk, which has important implications for future intervention efforts.

Authors

• Sheri Madigan, University of Calgary (U of C)
  Presenting Author

• Pasco Fearon, University College London (UCL)
  Non-Presenting Author

• Anh Ly, University of Calgary
  Non-Presenting Author

• Jessica Cooke, University of Calgary (U of C)
  Non-Presenting Author

• Marije L. Verhage, Vrije Universiteit Amsterdam
  Non-Presenting Author

• Carlo Schuengel, Vrije Universiteit Amsterdam
  Non-Presenting Author

• Robbie Duschinsky, University of Cambridge
  Non-Presenting Author

• Mirjam Oosterman, Vrije Universiteit Amsterdam
  Non-Presenting Author

• Marian J Bakermans-Kranenburg, ISPA Lisbon
  Non-Presenting Author

• Marinus H. van IJzendoorn, University College London (UCL)
  Non-Presenting Author