Adaptation to Quechua Collao and psychometric analysis of the instrument for the detection of violence against women

Abstract

Objectives. To determine the validity, measurement invariance, and reliability of the Quechua Collao version of the Woman Abuse Screening Tool (WAST) in Peruvian  women. Materials and methods: This was a psychometric study aimed at adapting the WAST to Quechua Collao as spoken in the Puno and Cusco regions, using a non-probabilistic sample of 521 women, most of whom (240; 46.1%) were between 18 and 34 years old. The process began with forward and backward translation. Then, five expert judges reviewed the Quechua Collao version, and a focus group of women confirmed the clarity and  comprehension of the items. Confirmatory Factor Analysis (CFA), multigroup CFA for measurement  invariance, reliability analysis, and external validity  testing using Spearman’s Rho were subsequently  conducted. Results. A unidimensional model of the WAST Quechua Collao showed adequate goodness-of-fit indices (CFI = 0.995; TLI = 0.992; SRMR = 0.051; RMSEA = 0.063) and demonstrated measurement invariance  across age, educational level, place of residence, and monthly family income (ΔCFI or ΔRMSEA < 0.01).  Regarding external validity, the WAST Quechua Collao  showed a moderate positive correlation with the Patient  Health Questionnaire (PHQ-9) (Rho = 0.618, p = 0.000). It also demonstrated excellent reliability, with Cronbach’s alpha (ɑ) = 0.860 and McDonald’s omega (ω) = 0.867.  Conclusions. The unidimensional WAST Quechua Collao version demonstrated internal structural validity, external validity, measurement invariance across age, educational level, place of residence, and monthly family income, as well as excellent reliability. Its use is recommended for screening intimate partner violence among Quechua Collao–speaking women.