Heterogeneous Effects of Parental Spanking on Subsequent Child Protective Services Involvement

Purpose:

Child maltreatment is a major public health concern that is associated with lasting negative consequences on child wellbeing.  Recent estimates indicate that more than 3 million children in the U.S. were referred to Child Protective Services (CPS) for potential maltreatment.  Accumulating research finds evidence that spanking, a legitimate and widely endorsed parenting practice in the U.S., is associated with higher risk of child maltreatment.  Despite compelling evidence on child and parent characteristics that are associated with spanking, scarce research considers the differences in these factors that may be subject to selection bias.  This study uses propensity score stratification method to estimate the possible heterogeneity in the effects of spanking on child maltreatment, while accounting for selection bias.  

Methods:

This study utilizes data from three waves of the Fragile Families and Child Wellbeing Study, a nationally representative study of urban families (n= 2,164).  At Wave 4 (age 5), mothers reported whether CPS contacted the family after the Wave 3 interview (yes/no).  At Wave 3 (age 3), spanking assessed whether the mother, father, or mother’s partner spanked the child in the past month (yes/no).  In the first step of our analysis, predictors at Wave 2 when the child was age 1 (i.e., child emotionality, maternal depression, substance use, demographics) estimated the propensity of parental spanking at age 3.  Based on their propensity score, families are stratified into three stratum and the associations of propensity of spanking at age 3 with subsequent CPS involvement in the next 2 years were estimated for all three stratum.  Subsequently, we examined whether child and parent characteristics that are possible predictors of CPS involvement were significantly different between the lowest and highest strata.

Results:

Findings demonstrate that children with the lowest propensity of spanking at age 3 (stratum 1) were at highest risk of CPS involvement during the subsequent 2 years (stratum 1: β = 0.88, p < .05; stratum 2: β = 0.44, p = .15; stratum 3: β = –0.10, p= .65).  Children with the lowest probability of spanking (stratum 1) were more likely to be females and have lower scores on negative emotionality than their counterparts with the highest propensity of spanking (stratum 3).  Also, children in stratum 1 had mothers who were older, more educated, in more stable relationships, less depressed, and who used less substances relative to children in stratum 3.

Conclusions and Implications:

These findings suggest that families characterized as least likely to spank were most likely to become involved in CPS in the actual event of spanking.  In research and practice, these families are considered the low risk group, and hence, their needs and potential vulnerability are easily overlooked.  This study highlights that these low risk families merit equal attention to prevent physical punishment that could lead to child maltreatment.  Additionally, families most likely to engage in spanking concurrently experience multiple risk factors such as depression and substance abuse, thus require a more comprehensive approach to prevent child maltreatment.