Psychometric evaluation of the WAST in the Collao Quechua: an integrative multi-method approach based on CFA, IRT, and network analysis

Abstract Background Violence against women is a problem with high prevalence in rural areas of Peru. Faced with this, the objective was to evaluate the psychometric properties of the Woman Abuse Screening Tool (WAST) in its Quechua version, Collao variant, under a multi-method approach. Method From a psychometric approach, an instrumental study was carried out in which 300 Quechua-speaking women from the Puno region participated and were administered the Collao Quechua version of the WAST, a scale of 8 items grouped into a single factor. The psychometric analysis included the application of Exploratory Graph Analysis (EGA) to examine dimensionality and structural stability, supplemented by a Confirmatory Factor Analysis (CFA) and a graded response model based on Item Response Theory (IRT). Reliability was estimated using the Omega coefficient and the test's information function. Result The EGA identified a highly stable one-dimensional structure (structural consistency = 1). The CFA showed an adequate fit (CFI = 0.985; TLI = 0.979; SRMR = 0.062), with high factor loadings (λ = 0.741-0.944) and high reliability (ω = 0.935). From the IRT, the model presented good global fit (CFI = 0.971; TLI = 0.935), adequate item discrimination, ordered thresholds, and absence of relevant local dependence. Tool accuracy was highest at medium to moderately high levels of the latent trait (θ ≈ 0 to 2), with high empirical reliability (rxx = 0.899). Conclusion The Collao Quechua version of the WAST presents robust psychometric properties, with high reliability and adequate functioning of its items. These findings support its validity as a screening instrument for the detection of violence in Quechua-speaking women.