The three factor model did exhibit configural invariance across men and women when sex was modeled as a grouping variable and compared to the three-factor model without this grouping variable, ΔX2(62) = 14.74, p = .999. However, the three factor model did not exhibit metric invariance across men and women when factor loadings were fit with equality constraints and this model was compared against the configural invariance model, ΔX2(10) = 249.63, p < .001. Indeed, the three-factor model exhibited a modestly better fit for men, X2 = 260.09, than women, X2 = 263.48. A potential reason for this differential model fit may be that the latent factor associations between two out of the three latent aggression factors were stronger for men than women (Table 6).
