Predictors of Sibling Violence: A Secondary Data Analysis of the Developmental Victimization Survey

Background and Purpose: Despite the high prevalence of sibling violence (Straus, Gelles, Steinmetz, 2006), this phenomenon has received minimal attention in family violence research. Most of the research on violence between siblings has focused on consequences individuals experience as a result of sibling violence (see Caffaro & Conn-Caffaro, 2005; Wiehe 1997) with less focus being placed on factors predicting its occurrence. The Developmental Victimization Survey conducted between 2002-2003 contains data on sibling violence, yet most of the research analyzing data from this survey has minimally considered sibling violence (see Finkelhor, Turner, and Ormrod, 2006). This poster will discuss an analysis of the Developmental Victimization Survey data to explore factors thought to be predictive of sibling violence.

Methods: Data from the Developmental Victimization Survey, which included the Juvenile Victimization Questionnaire (JVQ; Hamby, Finkelhor, Ormrod, & Turner, 2004), was analyzed to explore factors associated with predicting victimization by a sibling. Assault by a sibling is defined in the JVQ as including assault with or without a weapon, kidnapping, a biased attack, physical abuse by a sibling caregiver, gang or group assault with sibling perpetrator, and nonsexual genital assault by a sibling. The sample included 2030 respondents (1000 children aged 10-17 and 1030 caregivers of children aged 2-9) and is a nationally representative sample. Logistic regression was used to analyze the data; variables in the model included age group, gender, race/ethnicity, household income, parental education, parental supervision, parental criticism, child experience of physical abuse, child experience of psychological/emotional abuse, and child neglect. 

Results: Of the data on 2030 children, 1303 cases had data without missing values for any of the variables included. The Hosmer-Lemeshow Goodness-of-Fit Test, which included the previously mentioned variables, was not significant (χ²=7.701, df=8, p=.463) suggesting the model fit the data well. The following variables were significant predictors of sibling violence (odds ratio [95% CI]): age group (age 10-17 vs. ages 2-9; .61 [.47-.78], p<.001), child experience of psychological/emotional abuse (2.31 [1.58-3.37], p<.001), race/ethnicity of child (Non-Caucasian vs. Caucasian; .59 [.42-.84], p<.003), and parental supervision (.89 [.80-.99], p<.025). No other variables examined were statistically significant.

Conclusions and Implications: The findings of the data analysis indicate that age, race/ethnicity, parental supervision, and experience of parent-to-child emotional abuse are predictors of sibling violence. It was also found that sibling violence decreases with age and that Caucasian children are more likely to experience sibling violence; however, the results related to race/ethnicity need to be explored further as 81% of the sample was Caucasian. The results also suggest that sibling violence can occur regardless of household income and parental education meaning sibling violence may transcend socioeconomic boundaries. Further examination of the predictors of sibling violence is warranted. A better understanding of these predictors will help social workers create interventions aimed at addressing this form of family violence. Furthermore, knowledge of the predictors of violence between siblings can provide insight into ways in which social workers can implement preventive measures to circumvent this form of violence from occurring.