Measurement of Early Adolescent Dating Aggression during Middle School: Structure, Stability and Distinction from Peer Aggression

Introduction: Definitions of adolescent dating aggression (e.g., physical and psychological perpetration and victimization) have emerged from frameworks focusing adult intimate partner violence with researchers translating these constructs into measures for use with adolescents. However, this extension may be problematic for understanding adolescent dating aggression, particularly for the youngest daters (i.e., middle school students) where dating aggression may appear similar to peer aggression.  Further, it is not clear that measures of dating aggression developed for older adolescents and adults are appropriate for use with younger age groups. Developmental differences and the potential for dating aggression to look different across adolescent stages suggest the need to explore the relevance of dating aggression measures for younger age groups.  This paper will address these questions using two datasets.

Method: Study 1 evaluated the hypothesized structure of early adolescent dating aggression, and potential differences by gender and time from a high-risk sample of 2827 dating sixth grade students from four communities as part of the Multisite Violence Prevention Program. Study 2 replicated Study 1 findings in a second sample of 811 urban sixth, seventh and eighth grade dating adolescents from a large city in the Southeastern United States, while also exploring seasonal differences (i.e., Fall, Winter, Spring and Summer). In both studies, adolescent dating aggression was assessed by a modified version of the Safe Dates dating aggression measure (Foshee et al., 1996). Dating adolescents rated how frequently their boyfriend or girlfriend committed specific acts of dating aggression toward them and how frequently they did these things to their boyfriend or girlfriend.

Results: Compared to a two-factor model (perpetration, victimization) and a one factor model, a four-factor model comprised of physical perpetration, physical victimization, psychological perpetration and psychological victimization best fit the data in both Study 1 [CFI = 0.979, TLI = 0.975, RMSEA = 0.037] and Study 2 [CFI = 0.977, TLI = 0.972, RMSEA = 0.031]. Measurement invariance by gender, time and season was also examined.

Conclusion: The structure of dating aggression for early adolescents appears to resemble the structure borrowed from the adult literature. A further question remains as to whether these constructs are distinct from other forms of peer aggression common to this age. Therefore, discriminant validity analyses will also assess the extent to which dating aggression represents a distinct construct from other forms of peer aggression, namely physical and relational aggression perpetrated toward and received from peers.