Poster #60 - Secure Base Script Development in a non-Western Context: A 9-Year Longitudinal Study of Urban Chinese Families

• In Event: Poster Session 02

Abstract

One of the central propositions of attachment theory is that individuals internalize their early caregiving experiences as a generalized mental representation of attachment with its origins in early caregiving experience (Bowlby, 1973; 1980). Specifically, this representation is thought to originate from parental sensitivity/responsiveness to the child across childhood (i.e., the sensitivity hypothesis; van IJzendoorn & Sagi-Schwartz, 2008) and be generalized across different attachment relationships (e.g. mother/father). Interestingly, all long-term longitudinal tests of the generalization of attachment or the sensitivity hypothesis have been conducted in Western contexts. Combined with recent critiques of the universality of attachment theory from a cross-cultural perspective (Keller, 2018), the lack of longitudinal studies of attachment development outside the West is a significant weak point. In an effort to address this gap, we present the first longitudinal test of attachment generalization and the sensitivity hypothesis in a non-Western context based on a long-term prospective study of attachment in Nanjing, China.

Specifically, we analyzed data collected from a subsample of a 14-year longitudinal study of Chinese families (n = 356, 48.88% female). Lab observations of mother-infant interactions were collected at age 14 and 24 months. Maternal sensitivity was assessed and composited from 20-minute free play sessions which were coded by two trained and reliable coders using three 5-point scales: following-the-lead, intrusiveness, and positive regard (NICHD Early Child Care Research Network, 1999; ICCs = .79-.92; α = .76). Attachment security was operationalized as knowledge of the secure base script. Secure base script knowledge was assessed and composited at age 10-years using the middle childhood version of the attachment script assessment (ASA; Waters et al., 2015) which was adapted to the Chinese context (two coders, two mother and two father attachment narratives; ICCs = .86-.95; four stories, α = .77).

We examined generalization using Confirmatory Factor Analysis (CFA) to analyze the convergence of children’s secure base script knowledge for mother and father from the ASA (Waters et al., 2015). In the first step, we allowed the two latent factors to freely correlate, which aligns with the assumption that attachment security is generalized across intimate relationships (script knowledge for fathers and mothers, Figure 1). We also tested an alternative model where the correlation between the two domains was fixed to 0.01. When we allowed the two domains to freely correlate, model fit was good, χ2(1, N=251) = 2.115, p = .146, RMSEA = .067, CFI = .995, TLI = .972. The estimated standardized correlation between the two factors was .93. Model fit for the alternative model was significantly poorer (Δχ2 = 95.416, p<.001).

To test the sensitivity hypotheses, we employed multiple regression with FIML to examine the relations between maternal sensitivity during infancy/toddlerhood and attachment in middle childhood. After controlling for child’s biological sex and maternal education, maternal sensitivity significantly predicted child’s attachment at age 10 years, b = .22, S.E. = .09, β = .15, p = .031 (Table 1).

This longitudinal study provides the first support for the generalization of attachment and the sensitivity hypothesis in a non-Western context.

Authors

• Theodore Waters, New York University Abu Dhabi
  Presenting Author

• Rui Yang, New York University Abu Dhabi

• Yufei Gu, New York University Abu Dhabi

• Lixian Cui, NYU Shanghai

• Xuan Li, New York University

• Sumie Okazaki, New York University

• Xinyin Chen, University of Pennsylvania

• Hirokazu Yoshikawa, New York University

• Niobe Way, New York University