Measurement Invariance of the Reactive and Proactive Aggression Questionnaire (RPQ) across Self- and Other-reports

Abstract

ABSTRACT Many researchers have used the self-report version of the Reactive and Proactive Aggression Questionnaire (RPQ) to test aggression in other-reports in recent years. However, the validity and applicability of the self-report version of RPQ used as an other-report version have not been tested. This study aims to test the measurement invariance of the RPQ across self- and other-reports by multi-group confirmatory factor analysis. There were 485 students from grade 4 to grade 6 who completed the questionnaire by themselves, and 485 students from grade 1 to grade 6 who were rated by their mothers using the self-report version of RPQ. Item analysis using the graded response model (GRM) reveals 22 items (except item 19) have high discrimination and reasonable difficulty distribution. Confirmatory factor analyses were conducted on RPQ in a self-report sample and in an other-report sample. The results suggested that both samples fit two-factor model well (TLI =.933, CFI =.940, and RMSEA =.041 for self-report sample; TLI =.947, CFI =.952, RMSEA =.520 for other-report sample). Measurement invariance analysis for RPQ across self- and other-reports was tested after deleting item 22. The results support that the self-report version of RPQ is of high validity and invariance when it is used as an other-report instrument.