Academic Resilience After Maltreatment: Risk and Protective Networks Among Children Involved in Early Head Start

• In Event: 2-046 - Relationship of Child Outcomes to Child Maltreatment Types, Social Networks, and Early Head Start Programming

Integrative Statement

Introduction. Researchers increasingly recognize the importance of identifying protective factors associated with resilience to child maltreatment to inform successful intervention approaches. Little research has differentiated protective processes associated with resilient developmental outcomes—especially following maltreatment. This paper uses network analysis to examine whether children who have had at least one substantiated maltreatment report have different risk and protective ecologies than children who did not.
Sample. The network analysis included 2749 children randomly assigned to Early Head Start (EHS) or a control group in 16 of 17 EHS programs in 15 states included in the National Early Head Start National Research Project. Children were less than 1 year old at enrollment (1996-98). At the pre-kindergarten (Pre-K) assessment, children averaged five years old (63 months; Range = 54-76 months); 53% (n = 486) were male.

Method. Child welfare agencies matched study children with administrative child welfare records. Data on a set of 48 risk and protective factors and child outcomes were collected in the original EHS national study. Networks were estimated from correlation matrices (at 36 months); “edges” or links between “notes” or variables were defined as r >= .20). The outcome of interest was fifth grade language/literacy (median dichotomized). Maltreatment was defined as at least one substantiated report between the ages of 0 and 11 (prior to fifth grade outcome). Thus, we compared four networks: 1) no maltreatment, high language/literacy (NMHL); 2) no maltreatment report, low language/literacy (NMLL), 3) maltreatment, high language/literacy (MHL); and 4) maltreatment, low language/literacy (MLL).
Results and Conclusions. Network analyses revealed that children who had been maltreated had higher density networks (13% of possible edges) than those who had not been maltreated (8% of possible edges). Children who had been maltreated also had more connections within risk factors than those who had not (54 connections vs. 29 connections). To assess resilience from maltreatment, we compared MHL and MLL networks. The MLL network had more unique edges between child characteristics (e.g., temperament, vocabulary, emotional regulation) and stress/parenting difficulties (detached, dysfunctional interactions) than the MHL network (28 edges vs. 13 edges). The MHL network had more unique edges linking stress and parenting difficulties than the MLL network (43 edges vs. 29 edges), but only 7% of the edges connected stress/parenting difficulties to child characteristics (24% of the edges in the MLL network). Positive parenting characteristics had more unique links within the MHL network than the MLL network (18 edges vs. 9 edges, respectively). Findings suggest that child maltreatment may be the product of a higher degree of mutually-reinforcing risk factors (e.g., parent stress, homelessness and housing instability, lack of safety and warmth in home). The Law of Requisite Variety suggests that more interactive complexity (greater density) reduces a system’s ability to absorb, recover from, and adapt to adversity. Children’s resilience from maltreatment was associated with a decoupling of stress/parenting difficulties and children’s characteristics and a more pervasive influence of positive parenting characteristics.

Authors

• Rachel Chazan-Cohen, Tufts University
  Non-Presenting Author

• Carrie Furrer, Portland State University
  Presenting Author

• Teresa Schmidt, Portland State University
  Non-Presenting Author