Targeted Peer Victimization and the Construction of Positive and Negative Self-Cognitions: Connections to Depressive Symptoms in Children

Abstract

The goal was to examine the relation of covert/relational and overt/physical targeted peer victimization (TPV) to each other, to positive and negative self-cognitions, and to symptoms of depression. In a sample of elementary and middle school children, TPV was assessed by self-report, peer-nomination, and parent report in a multitrait–multimethod study. Positive and negative self-cognitions and depressive symptoms were assessed by self-report. Confirmatory factor analytic results support the convergent and discriminant validity of these two types of TPV. Both kinds of TPV were significantly related to positive and negative self-cognitions as well as self-reported depressive symptoms; however, structural equation modeling revealed that the effects of covert/relational TPV accounted for the effects of overt/physical TPV. In exploratory analyses, positive and negative self-cognitions explained the relation between TPV and depressive symptoms.

Notes

Note. Factor correlation was .44.

Note. Factor correlation = .21.

Note. RV = relational victimization; SR = self-report; PR = parent-report; PN = peer nomination; PV = physical victimization; CTI-C = Cognitive Triad Inventory for Children; CATS = Children's Automatic Thoughts Scale; SPPC = Self-Perception Profile for Children; CDI = Children's Depression Inventory.

Significance of gender mean differences: ∗∗p < .01. ∗∗∗p < .001.

¹We also tried other categorization criteria. Using more conservative criteria (naturally) diminished the overall rates of both kinds of TPV. However, the ratio of overt/physical to covert relational TPV remained very stable across a wide range of cutoffs.

²We also tested a model with Sex × Grade interaction term. However, this effect was not significant for either covert/relational or overt/physical TPV (ps > .20).

³For both sets of analyses (one predicting positive and negative cognitions, and the other predicting the five specific domains), we also tested the interaction between overt/physical and covert/relational TPV. Using Ping's (1996) two-step method, we found no evidence of an interaction between the two types of TPV in the prediction of any outcome variable (ps > .10).