Background: Extant literature has primarily employed linear models to estimate the average effect of spanking on children. Less is known about child and parent characteristics that may predict differential risks of children's exposure to spanking (i.e., pre-treatment heterogeneity) and the effect of spanking on child outcomes that may differ by the propensity for spanking (i.e., post-treatment heterogeneity).
Objective: The present study examines pre- and post-treatment heterogeneity in the association between maternal spanking at child age 3 and subsequent household Child Protective Services (CPS) involvement that were reported by mothers between child ages 3 to 5.
Methods: The sample consisted of 2,422 families from the Fragile Families and Child Wellbeing Study, a birth cohort study that oversampled children born to unmarried families. A propensity score stratification analysis procedure examined pre- and post-treatment heterogeneity in the association between spanking and subsequent household CPS involvement. Children were stratified into three strata depending on their propensity for experiencing spanking. Subsequently, the association between spanking and household CPS involvement was examined for each stratum.
Results: Disadvantaged socio-economic characteristics, negative psychosocial conditions of the mother, and dysregulated temperament of the child were associated with elevated risk of spanking. The odds of subsequent household CPS involvement increased by 2.4 times (β = 0.869, p <.01) for children in the stratum with the lowest propensity for experiencing spanking.
Conclusions: Results emphasize the importance of promoting alternative disciplinary strategies to families who would be considered low-risk before they may come into contact with the child welfare system.
Keywords: Child Protective Services involvement; Heterogeneous effects; Propensity score stratification analysis; Spanking.