Validation of a scale to measure ill-treatment of women

Abstract

Objectives. To analyse the validity of content and of structure and the reliability of a questionnaire designed to measure ill-treatment (IT) of women by their partners. Design. Descriptive, transversal, multi-centre study. Setting. Four urban health centres in Granada, Spain. Participants. Three hundred and ninety one women of 14 and over who consulted in primary care and had a stable partner for at least 3 months. Main measurements. Questionnaire by means of interviews between December 2000 and May 2001, with 10 Likert-like questions on physical, psychological, and sexual mistreatment, social and demographic questions and various health indicators. We analysed content validity by means of exploratory factorial analysis, reliability of alpha-Cronbach factors and of corrected scale-item correlation coefficients, and structure validity. Results. We obtained 2 empirical IT factors that did not correspond to the theoretical dimensions of physical, psychological, and sexual mistreatment and explained 64.21% of variance. The first included all the questions on psychological abuse, one on physical abuse and the sexual abuse dimension. The second covered the remaining questions on physical IT and "breaking things in the home". The factors gave reliability coefficients of 0.8688 and 0.7072. Comparison by means of extreme groups technique showed that the questionnaire's structure is valid. Conclusion. We found this was a reliable and valid questionnaire for evaluating ill-treatment of women. Its use, particularly in primary care, could help expand and deepen understanding of the problem.