Poster #161 - Validation of the Preschool and Early-Childhood Attachment Rating Scales, a Continuous Measure of Preschool Attachment

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Parenting & Parent-Child Relationships

Integrative Statement

Attachment theory provides a theoretical framework to understand how children’s early interactions with their caregivers come to shape their subsequent development. A notable recent focus of attachment research is aimed at understanding the specificity of preschool attachment, an important developmental period for children’s socialization and socioemotional development. Qualitative differences in child-caregiver preschool attachment are commonly coded with the MacArthur Preschool Attachment Coding System (PACS; Cassidy et al., 1992), which uses a categorical approach, similar to the infant Strange Situation coding system (Ainsworth et al., 1978). However, categorical assessments limit findings in three main regards: (a) they reduce statistical power, often forcing secure-insecure comparisons; (b) they run counter to evidence suggesting that infant and adult attachment are dimensionally distributed, as opposed to categorically defined (e.g., Fraley & Spieker, 2003) and; (c) they imply that children present negligible levels of behaviours associated with other classifications. In order to palliate to these limitations, this study aims to establish reliability and validity with child-mother dyads for a new continuous attachment system, the Preschool and Early-Childhood Attachment Rating Scales (PARS; Moss et al., 2015). This instrument comprises a 1-9 scale for each preschool attachment type: security, avoidance, ambivalence, behavioral-disorganization, caregiving-control, and punitive-control. We expected these scales to present adequate reliability, and content validity (negative associations between security and all insecure scales, and positive associations between behavioral-disorganization, caregiving-control, and punitive-control; Moss et al., 2015). We also expected to find criterion validity (positive association between security scale and maternal sensitivity; De Wolff & van IJzendoorn, 1997) and divergent validity (low to non-significant associations with child temperament; Groh et al., 2017). Lastly, we expected to find convergent validity (positive associations between each scale and their contiguous classification). One-hundred-and-forty-four children (83 girls) between the ages of 3 and 5 (M=3.91, SD=.73) participated in a laboratory session with their mother. Children’s attachment behaviors in the separation-reunion procedure (Cassidy et al., 1992) were coded by two independent teams using the PARS (Moss et al., 2015) and the MacArthur Preschool Attachment Coding System (PACS; Cassidy et al., 1992). We also assessed maternal sensitivity through a dyadic Laughing Task Procedure (Bureau et al., 2014) and child temperament through the Child Behavior Questionnaire (Putnam & Rothbart, 2006). Results attest of the PARS’ validity (see intraclass correlations in Table 1) and validity. Indeed, findings reveal negative correlations between the security scale and all insecure scales. Behavioral-disorganization, caregiving-control, and punitive-control scales were also significantly correlated. Furthermore, the PARS presented significant associations with maternal sensitivity (positive with security, negative with ambivalence and behavioral-disorganization), but none with child temperament. Lastly, a MANOVA revealed that children classified into a given category presented higher scores on the associated scale than their counterparts (e.g., secure children had significantly higher security scores; see Table 2). These results support the reliability and validity of the PARS to measure child-mother attachment in the preschool years. They also provide empirical support to the theorized underlying spectrum of disorganized child-mother attachment (O’Connor et al., 2011). Findings thereby support the application of continuous assessments of child-caregiver attachment.

Authors

• Audrey-Ann Deneault, University of Ottawa
  Presenting Author

• Kim Yurkowski, University of Ottawa
  Non-Presenting Author

• Ellen Moss, Université du Québec à Montréal
  Non-Presenting Author

• Jean-Francois Bureau, University of Ottawa
  Non-Presenting Author