Internal Factor Structure and Convergent Validity of a Perceived Partner Violence Scale in Peruvian Adults: Bifactor Model of Emotional, Physical, and Sexual Abuse and Sexual Manipulation

Abstract

Intimate partner violence is a social problem that requires appropriate instruments for its correct and accurate assessment. The aim of this study was to construct and report evidence of the psychometric properties of the Perceived Partner Violence Scale. Four samples were selected using nonprobability sampling. An exploratory factor analysis (EFA) was performed with the first sample. With the second sample, a confirmatory factor analysis (CFA) and validity testing based on the relationship with loneliness and self-esteem were carried out. With the third sample, validity was assessed based on the relationship with emotional dependence. With the fourth sample, validity was analyzed in relation to love compatibility. Robust methods such as the robust maximum likelihood method were used for the hierarchical models (bifactor model and second-order model), and diagonally weighted least squareswere used for the nonhierarchical models (correlated factor model, uncorrelated factor model, and unidimensional model). Validity tests report adequate fit indices in the EFA and CFA for a four-factor model (emotional violence, physical violence, sexual abuse, and sexual manipulation) as well as validity based on the relationship with other variables (self-esteem, loneliness, emotional dependence, and love compatibility). Robust alpha and omega estimators indicate acceptable internal consistency.