Structural equation modelling in male sexual violence against women research: best practices in statistical reporting and interpretation

Structural equation modelling (SEM), which is a family of statistical techniques in multivariate analysis, is increasingly used in sexual violence perpetration research. SEM is a versatile tool that allows for the examination of several latent variables in one cohesive analysis rather than conducting a series of sequential analyses. This advantage is of particular importance to sexual violence perpetration researchers, who often use several scales simultaneously to measure latent risk factors and outcome variables within one study. The current systematic review coded SEM statistical reporting practices in 28 sexual violence perpetration articles. We discussed statistical reporting trends across these studies and recommendations for best practice around SEM reporting techniques in sexual violence perpetration research. Researchers using SEM should be informed of best practices for reporting and interpreting SEM models to ensure adequate transparency in research communication, particularly given replicability crisis in social sciences, and the emergence of open science practices.PRACTICE IMPACT STATEMENTThis review identifies several trends in the reporting of structural equation modelling in sexual violence perpetration research. By analysing these trends alongside known best practices for reporting, we make recommendations to help future researchers design their studies and evaluate theoretical models of sexual violence perpetration using structural equation modelling.