Associations of Exposure to Neighborhood Disorganization and Maternal Spanking with Children's Aggression Using Fixed-Effects Models

Purpose:  The constellation of negative neighborhood and parent influences in childhood is a major public health concern that is linked to long-term risks throughout the life span.  Using the ecological framework, research provides solid evidence concerning the inter-relationships between neighborhood disadvantage, punitive parenting practices, and adverse child outcomes.  However, a recurrent debate is the extent to which the observed relationships between neighborhood, parenting, and child behavior are attributable to other unobserved confounding variables.  The present study examines the joint effects of neighborhood disorganization and maternal spanking using fixed effects regression for longitudinal data, a statistically rigorous approach that addresses many of the limitations in prior research.

Methods:  Data were obtained from the Fragile Families and Child Wellbeing Study, a birth cohort study in 20 U.S. cities. The present analysis was based on 2,472 families who participated in Wave 3 (child age 3) and Wave 4 (child age 5).  Mothers reported their child’s aggression using the Child Behavior Checklist.  Neighborhood disorganization was represented by three constructs that were mother’s self-report or interviewer observation: (1) collective efficacy, (2) crime and violence, and (3) physical disorder in neighborhood.  Frequency of maternal spanking in the past year was asked of mothers using a question from the Conflict Tactics Scale.  Covariates included maternal warmth, maternal depression, and demographics of mother and child.  Fixed-effects regression was employed to examine changes in aggression within the same child over time, while accounting for the effects of all unmeasured confounders that are consistent over time, such as initial level of child aggression, time-invariant genetic characteristics that may have associations with the use of parental spanking, as well as the selection to reside in certain neighborhoods. 

Results:  Results from fixed-effects regression indicated that collective efficacy had a significant inverse relationship with aggressive behavior of children (β = –0.032, p < .001), after accounting for other predictors and all time-invariant characteristics of the child and family that may have confounding relationships with the outcome.  Crime and violence in neighborhood was a significant predictor of aggressive behavior net of all other predictors in the model (β = 0.099, p < .001), whereas neighborhood physical disorder was not associated with aggression.  Children who were spanked by the mothers in the past year, even if it happened once or twice, had scored higher on aggression compared to their counterparts who were never spanked in the past year (β = 0.057, p < .001). 

Conclusions and Implications:  Our findings strongly support the importance of multilevel interventions that intervene at the level of both the family and the larger community for positive child outcomes.  Importantly, findings from this research suggest that there are separate pathways from both the neighborhood context, and parenting behavior to child outcomes.  Thus, programs and policy that promote child well-being would need to address risk factors at both neighborhood level and family level.  Furthermore, our results suggest the need for multifaceted neighborhood assessment that simultaneously considers structural and processes-oriented aspects of neighborhood conditions in programs that serve young children with behavioral issues.